CN110236558A - Method, device, storage medium and electronic equipment for predicting infant development - Google Patents

Abstract

Disclosed are a method, a device, a storage medium and electronic equipment for predicting a baby's developmental situation, belonging to the technical field of computer programs. The method includes: obtaining single limb movement data of the sample to be tested and corresponding month-age data; performing data processing to obtain a characteristic value corresponding to the month-age data for a single limb movement of the sample to be tested; The eigenvalue of the single limb movement corresponding to the sample to be tested and the corresponding month-old data are brought into the mathematical model of the stacked extreme learning machine for logical operation; the output of the logical operation of the stacked extreme learning machine is used as the test Based on the normal development or developmental delay of the sample, the prediction result of the baby's development is obtained. The apparatus, storage medium and electronic equipment can be used to implement the method. Through it, it can predict the development of the baby, and further obtain the possibility of the baby's developmental delay, and accordingly, it can provide a reliable basis for the early intervention of the baby's developmental delay.

Description

Method, device, storage medium and electronic equipment for predicting infant development

technical field

The present invention relates to the technical field of computer programs, in particular to a method, device, storage medium and electronic equipment for predicting infant development.

Background technique

In the prior art, it is usually necessary to wait until the infant is 2 years old to determine whether the infant is stunted. However, developmental delay in infants and young children was only diagnosed at this time, and the intervention for infants and young children was relatively late, which was not conducive to the recovery of infants and young children with developmental delay. Therefore, there is a need for a method that can predict the development of infants, so as to provide a basis for intervention in the recovery of infants with developmental delay.

Contents of the invention

In view of this, the present invention provides a method, device, storage medium and electronic equipment for predicting the development of a baby, through which the baby's development can be predicted, and the possibility of the baby's developmental delay can be further obtained. Delayed early intervention provides a reliable basis and thus more suitable for practical use.

In order to achieve the above-mentioned first object, the technical scheme of the method for predicting infant development provided by the present invention is as follows:

The method for predicting infant development provided by the invention comprises the following steps:

Obtain the single limb movement data and corresponding monthly age data of the sample to be tested;

Perform data processing on the single limb movement data of the sample to be tested, and obtain the eigenvalue corresponding to the monthly age data for the single limb movement of the sample to be tested;

Bringing a predetermined number of eigenvalues of a single limb movement corresponding to the sample to be tested and the corresponding monthly age data into the mathematical model of the stacked extreme learning machine to perform logical operations;

The output result of the logical operation of the stacked extreme learning machine is used as the basis for the normal development or developmental delay of the sample to be tested, and the prediction result of the baby's development is obtained.

The method for predicting infant development provided by the present invention can also be further realized by adopting the following technical measures.

As preferably, the construction method of the mathematical model of the stacked extreme learning machine comprises the following steps:

Obtain single limb movement data and corresponding monthly age data of the training sample, wherein the training sample is a sample of known development status;

Performing data processing on the single limb movement data of the training sample to obtain the eigenvalue corresponding to the monthly age data for the single limb movement of the training sample;

The eigenvalues corresponding to the month-age data for the single limb movement of the training sample are brought into the logical operation formula: When there are multiple training samples, multiple expressions based on the logical operation formula are obtained;

in,

L- is the number of nodes of the single hidden layer neural network,

g(x)-is the activation function,

w _i ＝[W _i1 ,w _i2 ,...,w _in ] ^T - is the input weight of the i-th hidden layer unit,

bi- is the bias of the i-th hidden layer unit,

β _i ＝[β _i1 ,β _i2 ,...,β _im ] ^T - is the output weight of the i-th hidden layer unit,

oj- is the logical operation output result after classification by stacking extreme learning machine for the training samples, wherein the logical operation output result includes two categories: normal development and developmental delay;

According to the multiple expressions based on the logical operation formula, determine the expression of the activation function g(x), the input weight wi of the ith hidden layer unit, the input weight wi of the ith hidden layer unit Bias bi and the output weight βi of the ith hidden layer unit;

Then the expression of the determined activation function g(x), the input weight wi of the i-th hidden layer unit, the bias bi of the i-th hidden layer unit, and the i-th hidden layer unit's The output weight βi is brought back to the logic operation formula to obtain the mathematical model of the stacked extreme learning machine.

Preferably, the single limb movement data includes:

The start time T _i0 , end time T _i1 , average acceleration a _v , peak acceleration _am , left leg motion type _SL , right leg motion type S _R ;

Wherein, the duration t _i of a single exercise is the difference T _i1 -T _i0 between the end time of a single exercise and the start time T _i0 of a single exercise.

As preferably, the method of performing data processing on the single limb movement data of the sample to be tested, and obtaining the eigenvalue corresponding to the monthly age data for the single limb movement of the sample to be tested is selected from univariate feature selection, recursive One of feature selection and stepwise feature selection, the data processing is performed on the single limb movement data of the sample to be tested, and the eigenvalue corresponding to the month-age data is obtained for the single limb movement of the sample to be tested Specifically include:

The univariate feature selection is selected as the feature value of the single limb movement through the statistical measurement method of the single variable in the single limb movement data; or

The recursive feature selection is performed by performing normalized data processing on each variable in the single limb movement data, and using the obtained data as the feature value of the single limb movement; or

The step-by-step feature selection selects a single variable in the single limb movement data one by one, and performs normalized data processing on each variable in the single limb movement data to obtain the data sequentially as the feature value of the single limb movement .

Preferably, in the output result of the logic operation,

According to the mathematical model of the stacked extreme learning machine, different classifications are carried out on the normal development;

Developmental delays are graded differently according to the mathematical model of the stacked extreme learning machine.

In order to achieve the above-mentioned second purpose, the technical scheme of the infant development situation prediction device provided by the present invention is as follows:

The infant development situation prediction device provided by the present invention comprises:

A data acquisition unit, configured to acquire single limb movement data and corresponding monthly age data of the sample to be tested;

A data processing unit, configured to perform data processing on the single limb movement data of the sample to be tested, and obtain the feature value corresponding to the monthly age data for the single limb movement of the sample to be tested;

The computing unit is used to bring the processed limb movement data corresponding to the sample to be tested and the corresponding monthly age data into the mathematical model of the stacked extreme learning machine for a predetermined number of times to perform logical operations;

The prediction result output unit is used to use the output result of the logical operation of the stacked extreme learning machine as the basis for the normal development or developmental delay of the sample to be tested to obtain the prediction result of the baby's development.

In order to achieve the above-mentioned third purpose, the technical solution of the storage medium provided by the present invention is as follows:

The storage medium provided by the present invention stores a baby development prediction program, and when the baby development prediction program is executed by a processor, the steps of the baby development prediction method provided by the present invention are realized.

In order to achieve the above-mentioned fourth purpose, the technical solution of the electronic equipment provided by the present invention is as follows:

The electronic equipment provided by the present invention includes a motion sensor, a processor, a memory, and a baby development prediction program stored on the memory and operable on the processor, wherein,

The motion sensor is used to obtain single limb movement data of the sample to be tested;

When the baby development prediction program is executed by the processor, the steps of the baby development prediction method provided by the present invention are realized.

The method, device, storage medium and electronic equipment provided by the present invention can obtain the single limb movement data and the corresponding monthly age data of the sample to be tested under the condition that the mathematical model of the stacked extreme learning system is known, that is, The eigenvalue of a single limb movement of the sample to be tested can be obtained through data processing, and the eigenvalue can be substituted into the known stacked extreme learning and mathematical model, and the output result can be obtained by performing logical operations, wherein the output result includes development Retarded and developmentally normal two categories, wherein, since the month-age data of the sample to be tested can be obtained according to the date of birth shown in the birth certificate, that is to say, in the baby development prediction method provided by the present invention, the data to be obtained is only It is a single limb movement data. In this case, it can be directly obtained by a motion sensor. Therefore, the application of the baby development prediction method provided by the invention can predict the baby development status more conveniently. In addition, the baby development status provided by the invention In the infant development prediction method, the stacked extreme learning and mathematical model can cover babies of all month ages during the construction process. Therefore, even if the sample to be tested is relatively young, the stacked extreme learning and The mathematical model obtained relatively accurate output results, therefore, it can provide a reliable basis for early intervention of infant developmental delay.

Description of drawings

Various other advantages and benefits will become apparent to those of ordinary skill in the art upon reading the following detailed description of the preferred embodiment. The drawings are only for the purpose of illustrating a preferred embodiment and are not to be considered as limiting the invention. Also throughout the drawings, the same reference numerals are used to designate the same parts. In the attached picture:

Fig. 1 is a schematic diagram of the structure of the baby development prediction equipment involved in the hardware operating environment of the embodiment of the present invention;

Fig. 2 is a flow chart of the steps of the method for predicting the development of an infant involved in the scheme of the embodiment of the present invention;

3 is a flow chart of the steps of the method for constructing the mathematical model of the stacked extreme learning machine used in the method for predicting the development of infants involved in the embodiment of the present invention;

Fig. 4 is a schematic diagram of the signal flow relationship among the functional modules in the baby development prediction device involved in the embodiment of the present invention;

Fig. 5 is a schematic diagram of the principle of the stacked extreme learning machine used in the method for predicting the development of infants involved in the solution of the embodiment of the present invention.

Detailed ways

In order to solve the problems existing in the prior art, the present invention provides a method, device, storage medium and electronic equipment for predicting the development of an infant, through which the development of the infant can be predicted, and the possibility of the infant’s developmental delay can be further obtained. According to this, It can provide a reliable basis for the early intervention of infant developmental delay, so it is more suitable for practical use.

In order to further explain the technical means and effects of the present invention to achieve the intended purpose of the invention, the following in conjunction with the accompanying drawings and preferred embodiments, the baby development prediction method, device, storage medium and electronic equipment proposed according to the present invention, its Specific embodiments, structures, features and effects thereof are described in detail below. In the following description, different "one embodiment" or "embodiment" do not necessarily refer to the same embodiment. Furthermore, the features, structures, or characteristics of one or more embodiments may be combined in any suitable manner.

The term "and/or" in this article is just an association relationship describing associated objects, which means that there can be three relationships, for example, A and/or B. The specific understanding is: A and B can be included at the same time, and A and B can be included separately. A exists, B may exist alone, and any of the above three situations can be met.

Referring to FIG. 1 , FIG. 1 is a schematic structural diagram of a device for predicting infant development in a hardware operating environment according to an embodiment of the present invention.

As shown in FIG. 1 , the baby development prediction device may include: a processor 1001 , such as a central processing unit (Central Processing Unit, CPU), a communication bus 1002 , a user interface 1003 , a network interface 1004 , and a memory 1005 . Wherein, the communication bus 1002 is used to realize connection and communication between these components. The user interface 1003 may include a display screen (Display), an input unit such as a keyboard (Keyboard), and the optional user interface 1003 may also include a standard wired interface and a wireless interface. The network interface 1004 may optionally include a standard wired interface and a wireless interface (such as a wireless fidelity (WIreless-FIdelity, WI-FI) interface). The memory 1005 may be a high-speed random access memory (Random Access Memory, RAM) memory, or a stable non-volatile memory (Non-Volatile Memory, NVM), such as a disk memory. Optionally, the memory 1005 may also be a storage device independent of the aforementioned processor 1001 .

Those skilled in the art can understand that the structure shown in Figure 1 does not constitute a limitation on the infant development prediction device, and may include more or less components than those shown in the illustration, or combine certain components, or arrange different components .

As shown in FIG. 1 , the memory 1005 as a storage medium may include an operating system, a data storage module, a network communication module, a user interface module, and a baby development prediction program.

In the baby development prediction device shown in Figure 1, the network interface 1004 is mainly used for data communication with the network server; the user interface 1003 is mainly used for data interaction with the user; the processor 1001 in the baby development prediction device of the present invention . The memory 1005 can be set in the baby development prediction device, and the baby development prediction device calls the baby development prediction program stored in the memory 1005 through the processor 1001, and executes the baby development prediction method provided by the embodiment of the present invention.

Embodiment one

Referring to accompanying drawing 1, the method for predicting the baby's development situation provided by Embodiment 1 of the present invention comprises the following steps:

Step 101: Obtain single limb movement data and corresponding monthly age data of the sample to be tested.

Specifically, here is to acquire the limb movement data of the sample to be tested through the sensor. Wherein, the limb movement data includes the duration of a single movement, the average acceleration a _v of a single movement, the peak acceleration a _m of a single movement, the movement type _SL of the left leg, and the movement type _SR of the right leg. Wherein, the single exercise duration t _i is the difference between the single exercise end time T _i1 and the single exercise start time T _i0 ie T _i1 −T _i0 . Among them, the left leg motion type S _L and the right leg motion type S _R can be displayed in the form of text description, for example, kicking, squatting, walking, crawling, bending, etc. can be qualitatively described, and a specific definition can be defined for each motion type. Symbols, such as kicking=Q ₁ , squatting=Q ₂ , walking=Q ₃ , crawling=Q ₄ , bending=Q ₅ , are then described symbolically. In addition, because infants develop their body movement ability more rapidly within 24 months, especially within 12 months, sometimes, even if the difference is only a few days, infants' body movements can undergo great changes. The data Y needs to be accurate to the day, for example, the month age data can be 2m+1, 8m+10, etc.

Step 102: Perform data processing on the single limb movement data of the sample to be tested, and obtain the feature value corresponding to the monthly age data for the single limb movement of the sample to be tested.

Specifically, the data processing method for the single limb movement data of the sample to be tested may include univariate feature selection, recursive feature selection, stepwise feature selection, etc., wherein,

The univariate feature selection refers to selecting the eigenvalue X _j of a single limb movement of the sample to be tested through some univariate statistical measurement methods. At this time, the first embodiment can be X _j ={T _i1 -T _i0 , m}; the second embodiment can be X _j ={a _v ,Y}; the third embodiment can be X _j ={a _m , Y}; in the fourth embodiment, X _j ={X _L , X _R , Y}.

Recursive feature selection refers to performing normalized data processing on the above-mentioned variables, and using the obtained data as the feature value X _j of a single limb movement of the sample to be tested. For example, at this time, X _j = {y ₁ (T _i1 -T _i0 )+y ₂ a _v +y ₃ a _m , X _L , X _R , Y}, where y ₁ is (T _i1 -T _i0 ) The characteristic coefficient, y ₂ is the characteristic coefficient of a _v , and y ₃ is the characteristic coefficient of a _m . In this embodiment, the normalized data processing method adopted is linear normalized data processing, and according to actual needs, some of the above three characteristics (T _i1 -T _i0 ), a _v , a _m can also be processed The square root operation can be performed on some of the above three features (T _i1 -T _i0 ), a _v , a _m to reduce the influence level.

Stepwise feature selection refers to selecting X _j = {T _i1 -T _i0 , Y} for the first time, selecting X _j = {a _v , Y} for the second time, and selecting X _j = {a _m , Y for the third time }, choose X _j ={X _L , X _R , Y} for the fourth time; then, choose X _j ={y ₁ (T _i1 -T _i0 )+y ₂ a _v +y ₃ a _m for the fifth time , X _L , X _R , Y}.

Step 103: Bring the predetermined number of eigenvalues of single limb movements corresponding to the sample to be tested and the corresponding month-age data into the mathematical model of the stacked extreme learning machine to perform logical operations.

Specifically, the eigenvalue X _j of the single limb movement of the sample to be tested is substituted into the mathematical model of the stacked extreme learning machine, namely

in,

L- is the number of nodes of the single hidden layer neural network,

g(x)-is the activation function,

w _i ＝[W _i1 ,w _i2 ,...,w _in ] ^T - is the input weight of the i-th hidden layer unit,

bi- is the bias of the i-th hidden layer unit,

β _i =[β _i1 ,β _i2 ,...,β _im ] ^T - is the output weight of the i-th hidden layer unit.

In the mathematical model of the trained stacked extreme learning machine, the expression of the activation function g(x), the input weight wi of the ith hidden layer unit, the bias bi of the ith hidden layer unit and The output weight βi of the i-th hidden layer unit is known, therefore, through the above logical operation, the output result oj can be obtained, which includes two categories of normal development and developmental delay.

Step 104: Using the output result of the logical operation of the stacked extreme learning machine as the basis for the normal development or developmental delay of the sample to be tested, the prediction result of the baby's development is obtained.

What needs to be explained here is that in the process of obtaining motion data for the same sample to be tested in the same time period, at least 5-10 groups of limb motion data need to be obtained before making a judgment. The reason is that if the sample to be tested is obtained When the number of limb movement data obtained is too small, it may be due to accidental factors that the movement of the sample to be tested is too intense or the movement is too slow, resulting in misjudgment. However, if within the same time period, at least When judging after obtaining 5-10 sets of limb movement data, it is possible to minimize judgment errors caused by accidental factors. If too many limb movement groups are obtained, it will take more time, and it is of little significance for judging whether the infant is developing normally or with developmental delay. If the sample to be tested is determined to be developmental delay after 5-10 sets of limb movement data, the number of limb movement groups can be increased by 1-3 times. Has a diagnostic role.

In the infant development prediction method provided by Embodiment 1 of the present invention, when the stacked extreme learning and mathematical model are known, the single limb movement data and the corresponding monthly age data of the sample to be tested can be obtained through data processing. The eigenvalue of a single limb movement of the sample to be tested is substituted into the known mathematical model of stacked extreme learning and logical operation, and the output results can be obtained. The output results include developmental delay and normal development 2 A large category, wherein, since the month-age data of the sample to be tested can be obtained according to the date of birth shown in the birth certificate, that is to say, in the baby development prediction method provided by the present invention, the data to be obtained is only a single limb movement Data, in this case, can be obtained directly by motion sensor, therefore, application baby development situation prediction method provided by the present invention can predict baby development situation comparatively conveniently, in addition, in the baby development situation prediction method provided by the present invention In the process of constructing the mathematical model of stacked extreme learning, it can cover babies of all ages. Therefore, even if the sample to be tested is young, it can be obtained through the mathematical model of stacked extreme learning. Accurate output results, therefore, it can provide a reliable basis for early intervention of infant developmental delay.

Referring to accompanying drawing 3 and accompanying drawing 5, the construction method of the mathematical model of stacked extreme learning machine comprises the following steps:

Step 201: Obtain single limb movement data and corresponding monthly age data of a training sample, where the training sample is a sample with known development status.

Specifically, here is to obtain the limb movement data of the training sample through the sensor. Wherein, the limb movement data includes the duration of a single movement, the average acceleration a _v of a single movement, the peak acceleration a _m of a single movement, the movement type _SL of the left leg, and the movement type _SR of the right leg. Wherein, the single exercise duration t _i is the difference between the single exercise end time T _i1 and the single exercise start time T _i0 ie T _i1 −T _i0 . Among them, the left leg motion type S _L and the right leg motion type S _R can be displayed in the form of text description. For example, kicking, squatting, walking, crawling, bending, etc. can be qualitatively described, and a specific definition can be defined for each motion type. Symbols, such as kicking=Q ₁ , squatting=Q ₂ , walking=Q ₃ , crawling=Q ₄ , bending=Q ₅ , are then described symbolically. In addition, because infants develop their body movement ability more rapidly within 24 months, especially within 12 months, sometimes, even if the difference is only a few days, the body movements of infants can undergo great changes. The data Y needs to be accurate to the day, for example, the month age data can be 2m+1, 8m+10, etc.

Step 202: Perform data processing on the single limb movement data of the training sample to obtain the feature value corresponding to the monthly age data for the single limb movement of the training sample.

Specifically, the data processing method for the single limb movement data of the training sample may include univariate feature selection, recursive feature selection, stepwise feature selection, etc., and the single limb movement data for the test sample The data is processed to obtain the eigenvalues corresponding to the month-age data for the single limb movement of the sample to be tested, which specifically include:

The univariate feature selection refers to selecting the eigenvalue X _j of a single limb movement as a training sample through some univariate statistical measurement methods. At this time, the first embodiment can be X _j ={T _i1 -T _i0 , m}; the second embodiment can be X _j ={a _v ,Y}; the third embodiment can be X _j ={a _m , Y}; in the fourth embodiment, X _j ={X _L , X _R , Y}. or

Recursive feature selection refers to performing normalized data processing on the above-mentioned variables, and using the obtained data as the feature value X _j of a single limb movement of the training sample. For example, at this time, X _j = {y ₁ (T _i1 -T _i0 )+y ₂ a _v +y ₃ a _m , X _L , X _R , Y}, where y ₁ is (T _i1 -T _i0 ) The characteristic coefficient, y ₂ is the characteristic coefficient of a _v , and y ₃ is the characteristic coefficient of a _m . In this embodiment, the normalized data processing method adopted is linear normalized data processing, and according to actual needs, some of the above three characteristics (T _i1 -T _i0 ), a _v , a _m can also be processed The square root operation can be performed on some of the above three features (T _i1 -T _i0 ), a _v , a _m to reduce the influence level. or

Stepwise feature selection refers to selecting X _j = {T _i1 -T _i0 , Y} for the first time, selecting X _j = {a _v , Y} for the second time, and selecting X _j = {a _m , Y for the third time }, choose X _j ={X _L , X _R , Y} for the fourth time; then, choose X _j ={y ₁ (T _i1 -T _i0 )+y ₂ a _v +y ₃ a _m for the fifth time , X _L , X _R , Y}.

Step 203: Bring the eigenvalues corresponding to the month-age data for a single limb movement of the training sample into the logical operation formula: When there are multiple training samples, multiple expressions based on logical operation formulas are obtained;

in,

L- is the number of nodes of the single hidden layer neural network,

g(x)-is the activation function,

w _i ＝[W _i1 ,w _i2 ,...,w _in ] ^T - is the input weight of the i-th hidden layer unit,

bi- is the bias of the i-th hidden layer unit,

β _i ＝[β _i1 ,β _i2 ,...,β _im ] ^T - is the output weight of the i-th hidden layer unit,

oj- is the output result of logical operation after classifying the training samples by stacking extreme learning machine. Among them, the output result of logical operation includes two categories: normal development and developmental delay.

What needs to be explained here is that in the case of training samples, the development of the training samples is known, that is, in the case of training samples, oj is known. Usually, oj includes two categories of developmental delay or normal development,

Step 204: Determine the expression of the activation function g(x), the input weight wi of the i-th hidden layer unit, the bias bi of the i-th hidden layer unit, and the i-th hidden layer unit according to a plurality of expressions based on logical operation formulas. The output weight βi of the hidden layer unit.

In this case, after the stacked extreme learning is completed, one only needs to input the eigenvalue X _j of the single limb movement of the sample to be tested to the stacked trace learning machine and determine the number of nodes L of the single hidden layer neural network , that is, the logical operation of oj can be output. That is to say, in this case, it is only necessary to obtain the characteristic value X _j of a single limb movement of the sample to be tested to determine the developmental status of the sample to be tested.

Step 205: The expression of the determined activation function g(x), the input weight wi of the i-th hidden layer unit, the bias bi of the i-th hidden layer unit, and the output weight βi of the i-th hidden layer unit Bring back to the logical operation formula to get the mathematical model of the stacked extreme learning machine.

Specifically, in this case, the mathematical model of the stacked extreme learning machine can be obtained as

in,

L - the number of nodes in the single hidden layer neural network,

g(x) - the activation function,

w _i ＝[W _i1 ,w _i2 ,...,w _in ] ^T - the input weight of the i-th hidden layer unit,

bi - the bias of the i-th hidden layer unit,

β _i =[β _i1 ,β _i2 ,...,β _im ] ^T - the output weight of the i-th hidden layer unit.

Among them, for the sample to be tested, the output result is unknown, according to the known expression of the activation function g(x), the input weight wi of the i-th hidden layer unit, and the bias of the i-th hidden layer unit Bi and the output weight βi of the i-th hidden layer unit and the number of nodes L of the single hidden layer neural network, output oj through logical operations.

Among them, the single body movement data includes: the start time T _i0 , the end time T _i1 , the average acceleration a _v , the peak acceleration _am , the left leg motion type _SL , and the right leg motion type _SR of the single body motion. Wherein, the duration t _i of a single exercise is the difference T _i1 -T _i0 between the end time of a single exercise and the start time T _i0 of a single exercise.

What needs to be explained here is that the reason why these indicators are selected as the data of a single limb movement is that the availability of motion sensor data is first considered. Then, the correlation between the data and the baby's development status should also be considered, among which the average acceleration a _v , peak acceleration _am , left leg motion type _SL , and right leg motion type S _R z can represent the baby's gross motor ability , therefore, they are used as the basic data for obtaining the eigenvalue X _j of a single limb movement. Wherein, the single limb movement data of the sample to be tested and the single limb movement data of the training sample are measured by the same motion sensor, or at least by sensors of the same type produced by the same manufacturer. The reason why the same sensor is at least the same type of sensor produced by the same manufacturer is because such a choice can reduce the judgment error caused by the sensor detection error.

Among them, the data processing is performed on the single limb movement data of the sample to be tested, and the method of obtaining the eigenvalue corresponding to the monthly age data for the single limb movement of the sample to be tested is selected from univariate feature selection, recursive feature selection, and stepwise feature selection. Choose one of them. in:

The univariate feature selection is selected as the feature value of a single limb movement through the statistical measurement method of a single variable in the single limb movement data.

Specifically, in this case, since the selected feature is a single variable, the calculation efficiency is high, but since the single variable usually considers fewer influencing factors, the possibility of error in the judgment result is relatively large .

Recursive feature selection processes the variables in the single limb movement data by normalizing data, and the obtained data is used as the feature value of the single limb movement.

Specifically, in this case, since the recursive feature selection comprehensively considers a single variable, and also considers the correlation between each single variable by referencing the characteristic coefficient, it can reduce the possibility of error in the judgment result .

Stepwise feature selection selects a single variable in the single limb movement data one by one, and performs normalized data processing on each variable in the single limb movement data to obtain the data as the feature value of the single limb movement in turn.

Specifically, in this case, combining single variables and normalized variables for comprehensive judgment can further reduce the possibility of errors in judgment results.

Among them, in the logic operation output result,

According to the mathematical model of the stacked extreme learning machine, the normal development is graded differently;

Developmental delays are graded differently based on a mathematical model of a stacked extreme learning machine.

Specifically, in some cases where specific subdivisions are required, more specific classifications can also be made for the classification of developmental delay and normal development, such as retardation-level 1, delay-level 2, delay-level 3, and delay-level 1 Grade 3 has the mildest degree of retardation, and Grade 3 has the most severe degree of retardation; normal-excellent, normal-good, normal-moderate. Specifically, on the basis of the judgment result output oj of the above-mentioned point variable feature selection, recursive feature selection, and stepwise feature selection, it is also possible to combine (T _i1 -T _i0 ), a _v , a _m to obtain the three detection data value and the categories of the two data of X _L and X _R , to further classify the normal or retarded, where, when oj is normal, the values of the three data (T _i1 -T _i0 ), a _v , a _m are larger , indicating that it is better when the development is normal. At this time, you can further set the corresponding thresholds for the three data of (T _i1 -T _i0 ), a _v , and a _m ; X _L , X The more categories of _R , the better it is under the condition of normal development. At this time, you can further set the corresponding thresholds for excellent, good, and medium for the categories of X _L and X _R. When oj is delayed, the smaller the three data values of (T _i1 -T _i0 ), a _v , and a _m are, the more serious the developmental delay is. At this time, it can be further targeted at (T _i1 -T _i0 ) The three data of , a _v , and a _m set the corresponding thresholds of level 1, level 2, and level 3 respectively; the fewer the categories of X _L and X _R , the more serious the developmental delay is. At this time, further For the categories of X _L and X _R , set the corresponding thresholds of level 1, level 2, and level 3 respectively.

Embodiment two

Referring to accompanying drawing 4, the device for predicting the baby's development situation provided by Embodiment 2 of the present invention includes:

The data acquisition unit 301 is used to acquire the single limb movement data and the corresponding monthly age data of the sample to be tested.

Specifically, here is to acquire the limb movement data of the sample to be tested through the sensor. Wherein, the limb movement data includes the duration of a single movement, the average acceleration a _v of a single movement, the peak acceleration a _m of a single movement, the movement type _SL of the left leg, and the movement type _SR of the right leg. Wherein, the single exercise duration t _i is the difference between the single exercise end time T _i1 and the single exercise start time T _i0 ie T _i1 −T _i0 . Among them, the left leg motion type S _L and the right leg motion type S _R can be displayed in the form of text description, for example, kicking, squatting, walking, crawling, bending, etc. can be qualitatively described, and a specific definition can be defined for each motion type. Symbols, such as kicking=Q ₁ , squatting=Q ₂ , walking=Q ₃ , crawling=Q ₄ , bending=Q ₅ , are then described symbolically. In addition, because infants develop their body movement ability more rapidly within 24 months, especially within 12 months, sometimes, even if the difference is only a few days, infants' body movements can undergo great changes. The data Y needs to be accurate to the day, for example, the month age data can be 2m+1, 8m+10, etc.

The data processing unit 302 is configured to perform data processing on the single limb movement data of the sample to be tested, and obtain the feature value corresponding to the monthly age data for the single limb movement of the sample to be tested.

Specifically, the data processing method for the single limb movement data of the sample to be tested may include univariate feature selection, recursive feature selection, stepwise feature selection, etc., wherein,

The univariate feature selection refers to selecting the eigenvalue X _j of a single limb movement of the sample to be tested through some univariate statistical measurement methods. At this time, the first embodiment can be X _j ={T _i1 -T _i0 , m}; the second embodiment can be X _j ={a _v ,Y}; the third embodiment can be X _j ={a _m , Y}; the fourth embodiment can be X _j ={X _L , X _R , Y};

Recursive feature selection refers to performing normalized data processing on the above-mentioned variables, and using the obtained data as the feature value X _j of a single limb movement of the sample to be tested. For example, at this time, X _j = {y ₁ (T _i1 -T _i0 )+y ₂ a _v +y ₃ a _m , X _L , X _R , Y}, where y ₁ is (T _i1 -T _i0 ) The characteristic coefficient, y ₂ is the characteristic coefficient of a _v , and y ₃ is the characteristic coefficient of a _m . In this embodiment, the normalized data processing method adopted is linear normalized data processing, and according to actual needs, some of the above three characteristics (T _i1 -T _i0 ), a _v , a _m can also be processed The square root operation can be performed on some of the above three features (T _i1 -T _i0 ), a _v , a _m to reduce the influence level.

Stepwise feature selection refers to selecting X _j = {T _i1 -T _i0 , Y} for the first time, selecting X _j = {a _v , Y} for the second time, and selecting X _j = {a _m , Y for the third time }, choose X _j ={X _L , X _R , Y} for the fourth time; then, choose X _j ={y ₁ (T _i1 -T _i0 )+y ₂ a _v +y ₃ a _m for the fifth time , X _L , X _R , Y}.

The computing unit 303 is used to bring the processed limb movement data corresponding to the sample to be tested and the corresponding monthly age data into the mathematical model of the stacked extreme learning machine for a logical operation.

Specifically, the eigenvalue X _j of the single limb movement of the sample to be tested is substituted into the mathematical model of the stacked extreme learning machine, namely

in,

L- is the number of nodes of the single hidden layer neural network,

g(x)-is the activation function,

w _i ＝[W _i1 ,w _i2 ,...,w _in ] ^T - is the input weight of the i-th hidden layer unit,

bi- is the bias of the i-th hidden layer unit,

β _i =[β _i1 ,β _i2 ,...,β _im ] ^T - is the output weight of the i-th hidden layer unit.

In the mathematical model of the trained stacked extreme learning machine, the expression of the activation function g(x), the input weight wi of the ith hidden layer unit, the bias bi of the ith hidden layer unit and The output weight βi of the i-th hidden layer unit is known, therefore, through the above logical operation, the output result oj can be obtained, which includes two categories of normal development and developmental delay.

The prediction result output unit 304 is used to use the output result of the logical operation of the stacked extreme learning machine as the basis for the normal development or developmental delay of the sample to be tested, so as to obtain the prediction result of the baby's development.

What needs to be explained here is that in the process of obtaining motion data for the same sample to be tested in the same time period, at least 5-10 groups of limb motion data need to be obtained before making a judgment. The reason is that if the sample to be tested is obtained When the number of limb movement data obtained is too small, it may be due to accidental factors that the movement of the sample to be tested is too intense or the movement is too slow, resulting in misjudgment. However, if within the same time period, at least When judging after obtaining 5-10 sets of limb movement data, it is possible to minimize judgment errors caused by accidental factors. If too many limb movement groups are obtained, it will take more time, and it is of little significance for judging whether the infant is developing normally or with developmental delay. If the sample to be tested is determined to be developmental delay after 5-10 sets of limb movement data, the number of limb movement groups can be increased by 1-3 times. Has a diagnostic role.

The infant development prediction device provided by Embodiment 2 of the present invention obtains the single limb movement data and the corresponding monthly age data of the sample to be tested under the condition that the stacked extreme learning and the mathematical model are known, which can be obtained through data processing. The eigenvalue of a single limb movement of the sample to be tested is substituted into the known mathematical model of stacked extreme learning and logical operation, and the output results can be obtained. The output results include developmental delay and normal development 2 A large category, wherein, since the month-age data of the sample to be tested can be obtained according to the date of birth shown in the birth certificate, that is to say, in the baby development prediction method provided by the present invention, the data to be obtained is only a single limb movement Data, in this case, can be obtained directly by motion sensor, therefore, application baby development situation prediction method provided by the present invention can predict baby development situation comparatively conveniently, in addition, in the baby development situation prediction method provided by the present invention In the process of constructing the mathematical model of stacked extreme learning, it can cover babies of all ages. Therefore, even if the sample to be tested is young, it can be obtained through the mathematical model of stacked extreme learning. Accurate output results, therefore, it can provide a reliable basis for early intervention of infant developmental delay.

Embodiment three

The storage medium provided by the third embodiment of the present invention stores the baby development prediction program, and when the baby development prediction program is executed by the processor, the steps of the baby development prediction method provided by the present invention are realized.

The storage medium provided by the third embodiment of the present invention can obtain the single limb movement data and the corresponding monthly age data of the sample to be tested under the condition that the mathematical model of stacked extreme learning is known, and the sample to be tested can be obtained through data processing The eigenvalue of a single limb movement is substituted into the known stacked extreme learning and mathematical model, and the output results can be obtained by performing logical operations. Among them, the output results include two categories: developmental delay and normal development. , wherein, since the month-age data of the sample to be tested can be obtained according to the date of birth displayed on the birth certificate, that is to say, in the baby development prediction method provided by the present invention, the data to be obtained is only a single limb movement data, in In this case, it can be directly obtained by the motion sensor. Therefore, the application of the infant development prediction method provided by the present invention can predict the infant development status more conveniently. In addition, in the infant development prediction method provided by the present invention, stacking In the process of constructing the mathematical model of stacked extreme learning, it can cover babies of all ages. Therefore, even if the sample to be tested is young, relatively accurate output can be obtained through the mathematical model of stacked extreme learning. Consequently, it can provide a reliable basis for early intervention of infant developmental delay.

Embodiment four

The electronic device provided by Embodiment 4 of the present invention includes a motion sensor, a processor, a memory, and a baby development prediction program stored in the memory and operable on the processor, wherein,

A motion sensor, used to acquire single limb movement data of the sample to be tested;

The steps of the infant development prediction method provided in Embodiment 1 of the present invention are realized when the infant development prediction program is executed by the processor.

The electronic device provided by Embodiment 4 of the present invention obtains the single limb movement data and the corresponding monthly age data of the sample to be tested under the condition that the mathematical model of the stacked extreme learning method is known, and the sample to be tested can be obtained through data processing The eigenvalue of a single limb movement is substituted into the known stacked extreme learning and mathematical model, and the output results can be obtained by performing logical operations. Among them, the output results include two categories: developmental delay and normal development. , wherein, since the month-age data of the sample to be tested can be obtained according to the date of birth displayed on the birth certificate, that is to say, in the baby development prediction method provided by the present invention, the data to be obtained is only a single limb movement data, in In this case, it can be directly obtained by the motion sensor. Therefore, the application of the infant development prediction method provided by the present invention can predict the infant development status more conveniently. In addition, in the infant development prediction method provided by the present invention, stacking In the process of constructing the mathematical model of stacked extreme learning, it can cover babies of all ages. Therefore, even if the sample to be tested is young, relatively accurate output can be obtained through the mathematical model of stacked extreme learning. Consequently, it can provide a reliable basis for early intervention of infant developmental delay.

While preferred embodiments of the invention have been described, additional changes and modifications to these embodiments can be made by those skilled in the art once the basic inventive concept is appreciated. Therefore, it is intended that the appended claims be construed to cover the preferred embodiment as well as all changes and modifications which fall within the scope of the invention.

Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and equivalent technologies thereof, the present invention also intends to include these modifications and variations.

Claims (8)

1. A method for predicting the development of an infant, comprising the steps of:

acquiring single limb movement data and corresponding month age data of a sample to be detected;

performing data processing on the single limb movement data of the sample to be detected to obtain a characteristic value corresponding to the month age data and aiming at the single limb movement of the sample to be detected;

substituting the feature value of single limb movement corresponding to the sample to be tested for a preset number of times and the corresponding monthly age data into a mathematical model of the stacked extreme learning machine, and performing logic operation;

and taking the output result of the logic operation of the stacked extreme learning machine as the basis of the normal development or the retarded development of the sample to be detected to obtain the prediction result of the development condition of the infant.

2. The method for predicting the development condition of an infant according to claim 1, wherein the method for constructing the mathematical model of the stacked extreme learning machine comprises the following steps:

acquiring single limb movement data and corresponding month age data of a training sample, wherein the training sample is a sample with a known development condition;

performing data processing on the single limb movement data of the training sample to obtain a characteristic value corresponding to the month age data and aiming at the single limb movement of the training sample;

substituting the characteristic value corresponding to the month age data and aiming at the single limb movement of the training sample into a logical operation formula:when the training samples are multiple, obtaining multiple expressions based on the logical operation formula;

wherein,

l is the number of nodes of the single hidden layer neural network,

g (x) is the activation function,

w_i＝[W_i1,w_i2,...,w_in]^Tis the input weight of the ith hidden layer unit,

bi is the bias of the ith hidden layer unit,

β_i＝[β_i1,β_i2,...,β_im]^Tthe output weight of the ith hidden layer unit,

oj is a logic operation output result after classification by applying a stacked extreme learning machine aiming at a training sample, wherein the logic operation output result comprises 2 major classes of normal development and delayed development;

determining β i an expression of the activation function g (x), an input weight wi of the i-th hidden layer unit, a bias bi of the i-th hidden layer unit and an output weight of the i-th hidden layer unit according to the plurality of expressions based on the logical operation formula;

and then bringing back the determined expression of the activation function g (x), the input weight wi of the ith hidden layer unit, the bias bi of the ith hidden layer unit and the output weight β i of the ith hidden layer unit to the logic operation formula to obtain the mathematical model of the stacked extreme learning machine.

3. The method of claim 1 or 2, wherein the single limb movement data comprises:

starting time T of single limb movement_i0End time T_i1Average acceleration a_vPeak acceleration a_mLeft leg type S_LRight leg type S_R；

Wherein the duration t of a single movement_iIs the end time of a single movement and the beginning time T of the single movement_i0Difference value T of_i1-T_i0。

4. The method for predicting the development condition of the infant according to claim 1 or 2, wherein the method for performing data processing on the single limb movement data of the sample to be tested to obtain the feature value of the single limb movement of the sample to be tested corresponding to the month age data is selected from one of univariate feature selection, recursive feature selection and step-by-step feature selection, and the step of performing data processing on the single limb movement data of the sample to be tested to obtain the feature value of the single limb movement of the sample to be tested corresponding to the month age data specifically includes:

selecting the univariate characteristics to obtain a characteristic value serving as the single limb movement through a statistical measurement method of the univariate in the single limb movement data; or

The recursive feature selection is to perform normalized data processing on all variables in the single limb movement data to obtain data serving as a feature value of the single limb movement; or

And the step-by-step feature selection is to select single variables in the single limb movement data one by one and to carry out normalization data processing on all the variables in the single limb movement data to obtain data which are sequentially used as feature values of the single limb movement.

5. The method of claim 2, wherein in the output of the logic operation,

according to the mathematical model of the stacked extreme learning machine, carrying out different grading on normal development;

and carrying out different grades on the developmental delay according to the mathematical model of the stacked extreme learning machine.

6. An infant development prediction apparatus comprising:

the data acquisition unit is used for acquiring single limb movement data and corresponding month age data of a sample to be detected;

the data processing unit is used for carrying out data processing on the single limb movement data of the sample to be detected to obtain a characteristic value corresponding to the month age data and aiming at the single limb movement of the sample to be detected;

the operation unit is used for bringing the processed limb movement data corresponding to the sample to be tested and the corresponding month age data of the preset times into a mathematical model of the stacked extreme learning machine for logic operation;

and the prediction result output unit is used for taking the output result of the logic operation of the stacked extreme learning machine as the basis of the normal development or the slow development of the sample to be tested to obtain the prediction result of the development condition of the infant.

7. A storage medium having a baby development prediction program stored thereon, wherein the baby development prediction program, when executed by a processor, implements the steps of the baby development prediction method according to any one of claims 1 to 5.

8. An electronic device comprising a motion sensor, a processor, a memory, and an infant development prediction program stored on the memory and executable on the processor, wherein,

the motion sensor is used for acquiring single limb motion data of a sample to be detected;

the infant development prediction program, when executed by the processor, implements the steps of the infant development prediction method according to any one of claims 1 to 5.