TWI402713B - A Method and System for Determining Binary Marks of Trust Region with Skew Potential Coefficient - Google Patents

Description

Method and system for determining binary target using trust interval of skew coefficient

The present invention relates to a method and system for determining a binary target, and more particularly to a method and system for determining a binary target using a confidence interval of a skew coefficient.

Hospital evaluation norms hospitals need to do a lot of investigations and analysis, such as physical examination of employees and analysis of medical examination report data, or monthly comparison of financial information accounting subjects and quality control indicators of various departments. These data analysis processes are often a heavy burden. Therefore, how to build an information module to help quickly complete the analysis report becomes an important challenge for information personnel.

However, many hospital questionnaire analysis workers do not have a deep statistical basis, so it is really troublesome to use the t or ANOVA test with the number of mothers and how to use the cumbersome statistical software and correctly interpret and explain the report. Analysts of many hospital questionnaires or binary target data comparisons. Therefore, it is necessary to seek a solution.

Accordingly, it is an object of the present invention to provide a method of determining a binary target using a confidence interval of a skew coefficient.

Therefore, the present invention uses the binary criterion determination method of the confidence interval of the skew coefficient, and is suitable for the statistical identification of the determination result of the binary target for the user. The binary target determination method comprises the following steps: (A) providing a set of binary target determination source data; (B) using a set of skew coefficient calculation module, randomly sampling the data from the set of binary target determination sources a predetermined number of sample samples, and calculating the skew coefficient from the sampled samples; (C) the skew coefficient calculation module repeats the step (B) until a second predetermined number is obtained a skew coefficient; (D) using the confidence interval calculation module to calculate a set of trust intervals of the skew coefficients from the second predetermined number of skew coefficients; and (E) using the graphical mode The group, graphically presenting the confidence interval of the skewed coefficients to represent the statistical analysis of the decision of the binary target required by the user.

Another object of the present invention is to provide a binary target decision system using a confidence interval of skew coefficients.

Therefore, the present invention uses a binary target determination system of the confidence interval of the skew coefficient, and is suitable for the user to determine whether the binary target is statistically significant. The binary target determination system comprises: a source data database, a skew coefficient calculation module, a trust interval calculation module and a graphic module. The source data database is used to provide a set of binary target decision source data. The skew coefficient calculation module is configured to randomly sample a first predetermined number of samples from the data of the set of binary target determination sources, and calculate a skew coefficient from the sampled samples, and then The steps of randomly sampling and calculating the skewness coefficient are repeated until the second predetermined number of skewness coefficients are obtained. The confidence interval calculation module is configured to calculate a set of trust intervals of the skew coefficients from the second predetermined number of skew coefficients. The graphic module graphically presents the confidence interval of the binary target of the skew coefficients to represent the determination result of the binary target required by the user.

The effect of the invention is that information personnel (such as hospital information personnel) can assist grassroots workers (such as hospital grassroots workers) to simplify the analysis steps, and can correctly and quickly analyze the data by topic (such as medical examination data, finance of various departments). Whether the accounting competition and quality indicator data are compared monthly (by comparison), whether the level of significant normal (or abnormal) is reached, or whether the questionnaire is analyzed to achieve a significant satisfaction (or dissatisfaction) level.

The foregoing and other technical features, features and advantages of the invention are apparent from the detailed description of the preferred embodiments.

Before the invention is described in detail, it is noted that in the following description, similar elements are denoted by the same reference numerals.

Referring to Figures 1 and 2, the present invention is directed to a binary index determination system 1 and method using a confidence interval of skewness (Skewness) for a user to obtain a binary target. The determination result of the binary target may be normal or abnormal of the data, for example, the normal or abnormal data of the medical examination report of the hospital employee. Alternatively, the determination result of the binary target may also be satisfaction or dissatisfaction of the satisfaction survey. In a preferred embodiment of the present invention, the developed information system utilizes the skewness coefficient of the data and combines techniques for statistically estimating the confidence interval (eg, 95% confidence interval) to assist hospital staff in accepting this. Many surveys (such as Likert's multi-point scoring model) and analysis. In the following, a physical examination report data (such as a two-point scoring mode of 1 and 0) of a 642 employee who participated in a health check in a regional hospital for a certain year will be taken as an example to describe a preferred embodiment of the present invention, wherein the medical examination item includes chest X-ray. , occult blood, triglycerides, blood sugar, creatinine, glutamic acid transaminating sputum (SGPT) and the like. For the normal and abnormal values of each test value, 1 and 0 are assigned respectively to perform the satisfaction (health) degree and difference analysis for each physical examination result. Using the graphical analysis method and system of the preferred embodiment of the present invention, the hospital information staff will be able to assist the primary hospital staff to simplify the analysis steps, and to correctly and quickly analyze whether the data reaches a significant satisfaction (or dissatisfaction) level. .

As shown in FIG. 1, the preferred embodiment of the binary index determination system 1 using the confidence interval of the skew coefficient includes a source data database 11, a skew coefficient calculation module 12, and a confidence interval calculation module. 13 and a graphic module 14. The steps of the method for determining the binary flag of the confidence interval using the skew coefficient of the present invention and the details of the elements of the binary target decision system 1 will be described below with reference to the flowchart of FIG.

First, as shown in step 21, a set of binary target decision source data 111 is first provided by the source data database 11, wherein the set of binary target decision source data 111 is generally referred to by the skew coefficient calculation module 12. The sampling is performed by the skew coefficient calculation module 12, and the original data of the skew coefficient is calculated therefrom. For example, in the preferred embodiment, the set of binary target determination source data 111 is derived from Tables 1 and 2 below, wherein Table 1 represents the basic demographic data of the 642 employees, and Table 2 represents the inspection project Lee. Kettle scale score classification.

In the preferred embodiment, the program module such as the skew coefficient calculation module 12 and the confidence interval calculation module 13 is written by VBA (Visual Basic for Application) in the Excel spreadsheet software. So far, the actual physical examination data of 642 employees based on Table 1 and Table 2 have been constructed in the form of Excel data sheets to become Tables 3 and 4. Table 3 represents the actual physical examination data sets of 642 employees (because the information is too large, only Three employee data and two X-ray and urine occult blood test items are listed as examples, and Table 4 represents a list of expected statistical analysis (listed only). In addition, the meanings of the digital annotations in each column of Table 3 are as follows: 0 in the gender column represents female, 1 represents male; 1 in the age column represents 30 years old, 2 represents 31-40 years old, and 3 represents 41- 50 years old, 4 representatives ≧ 51 years old; 1 representative in the seniority column ≦ 5 years, 2 representatives ≦ 10 years, 3 representatives ≦ 20 years, 4 representatives ≧ 21 years; 2 in the department column represent the administration, 3 representatives care 5 represents medical technology and 7 represents medical treatment.

Then, as shown in step 22 of FIG. 2, the skew coefficient calculation module 12 randomly samples a set of first predetermined number (such as 500) from the set of binary source determination source data 111, and from these samples. In the sample taken, a skew coefficient is calculated.

That is, the skew coefficient calculation module 12 performs sequential processing according to Table 4, for example, processing the data of all 642 subjects of the data set column 6 (chest X-ray) of Table 3, and then analyzing the second field group. And the notes are 0 (female) and 1 (male) column 6 (chest X-ray) items, and then the group ages 3, 4, 5, age, department, and then the column, then analyze the column 7 urine occult blood data, and so on. It should be particularly noted that the main function of the present invention is that the skew region can be automatically calculated from the determination source data of a set of binary targets by the skew coefficient calculation module 12 and the confidence interval calculation module 13. Then, the determination result of the binary target required by the user is presented by the graphic module 14. Therefore, in the process described above, all employee medical examination materials are divided into various groups, which are additional functions of the program module system 1 of the present invention. However, even if the system 1 does not have the function of the group automatic to table 3 data set crawling data divided by the table 4 processing list, the system does not have the function of dividing the binary target decision source data 111 into various groups. 1 is still within the scope of protection of the present invention.

In the preferred embodiment, the skew coefficient calculation module 12 performs random sampling of the sample data by using a shoe sampling method to obtain the first predetermined number of samples, for example, in Table 4. For example, if the first predetermined number is 500 samples, it means that the skew coefficient calculation module 12 uses the shoe sampling method from the trust interval of the chest X-ray examination item of the whole employee group. A total of 642 samples of 642 chest X-ray examination scores (normal 1 or abnormal 0) of 642 employees were randomly sampled.

As for the calculation of the skewness coefficient, the skew coefficient calculation module 12 calculates the skew coefficient y using the third-order motion difference formula of the following formula (1):

Where μ represents the average of the data set ( x _i ), σ is the standard deviation of x , E represents the expected value, y > 0 represents positive bias, and y < 0 represents negative bias.

Next, as shown in step 23 of FIG. 2, the skew coefficient calculation module 12 determines whether a second predetermined number (for example, 100) of skew coefficients has been obtained. In the embodiment of the present invention, since the confidence interval calculation module 13 has to obtain the trust interval of the plurality of groups (that is, the second predetermined number) of the skew coefficient, the skew coefficient calculation module 12 must repeat the steps. Random sampling of 22 to obtain the skew coefficient of the second predetermined number. That is, when the result of the determination in step 23 is no (ie, less than the predetermined 100 sets of skewness coefficients), the skew coefficient calculation module 12 proceeds to step 22 again. On the other hand, when the result of the determination in step 23 is YES, it indicates that the second predetermined number (for example, 100) of the skew coefficient has been obtained, and then proceeds to step 24.

As shown in step 24, after the skew coefficient calculation module 12 obtains the second predetermined number (for example, 100) of the skew coefficient, the confidence interval calculation module 13 then shifts from the second predetermined number. The state coefficient calculates the confidence interval of the skew coefficients and obtains the median of the skew coefficients, wherein the confidence interval can be, for example, a 95% confidence interval.

As shown in Table 5 below, the method for determining the binary target using the confidence interval of the skew coefficient of the present invention and the system 1 for the physical examination data of the 642 employees (represented by 1) or abnormal (represented by 0) ), the calculated median of the skewness coefficient and the 95% confidence interval and other related data, but as shown in Table 4, since there are many group categories divided in this embodiment, only Table 5 lists Some groups of data, for example for chest X-ray items, only list classification age-1 (representing ≦30 years old), age-4 (representing ≧51 years old), seniority-2 (representing ≦10 years) and seniority- 4 (representing ≧ 21 years) and other four categories of data.

Note: The confidence interval spacing indicates the difference between the confidence intervals of the binary targets. When it is greater than 0, it means that there is no overlap and there is a statistically significant difference, otherwise it overlaps without difference.

Referring to FIG. 3, then, as shown in FIG. 2 and FIG. 3, the graphic module 14 automatically displays the 95% confidence interval and the median data point of the skew coefficient in FIG. 3 in FIG. 3 . The user can determine whether the 95% confidence interval of the result of the normal or not of a certain type of physical examination data of a certain group is located on the left side or the right side of the zero skew coefficient axis 141. The results of this type of physical examination are statistically significant normal or abnormal. For example, from the X-ray project of the employee of age-1 (representing ≦30 years old) in the top row of Figure 3, since the 95% confidence interval is completely located on the left side of the zero skew coefficient axis 141, the user can It is known that the chest X-ray examination results of employees whose age is less than or equal to 30 years old should be statistically significant. For another example, from the X-ray project of the employee of the seniority-4 (representing ≧21) in the fourth column of Fig. 3, since the 95% confidence interval spans the left and right sides of the zero skew coefficient axis 141, it indicates The X-ray physical examination results of the employees whose annual salary is greater than or equal to 21 years are close to the normal distribution, rather than the statistically significant normal or abnormal physical examination results.

In summary, the method and system for determining the binary target of the confidence interval using the skew coefficient of the present invention utilizes the skew coefficient of the data and the technique of statistically estimating the confidence interval, so that information personnel (such as hospital information personnel) can assist the grassroots. Workers (such as hospital grassroots workers) streamline the analysis steps, and can correctly and quickly analyze whether the data reaches a significant normal (or abnormal) level, or whether the questionnaire is subject to a significant satisfaction (or dissatisfaction) level. Therefore, the object of the present invention can be achieved.

The above is only the preferred embodiment of the present invention, and the scope of the invention is not limited thereto, that is, the simple equivalent changes and modifications made by the scope of the invention and the description of the invention are All remain within the scope of the invention patent.

1. . . Binary target decision system using the confidence interval of skew coefficient

11. . . Source data database

111. . . Binary target decision source data

12. . . Skewed coefficient calculation module

13. . . Trust interval calculation module

14. . . Graphic module

141. . . Zero skew coefficient axis 141

21~25. . . step

1 is a system block diagram showing a preferred embodiment of the binary index determination system of the confidence interval using the skew coefficient of the present invention;

2 is a flow chart showing a preferred embodiment of the method for determining a binary target of a confidence interval using a skew coefficient of the present invention;

Figure 3 is a statistical diagram illustrating a statistical graph of skewed coefficient and median information generated by the system and method of the present invention for a binary target determination of the 642 hospital staff.

1. . . Binary target decision system using the confidence interval of skew coefficient

11. . . Source data database

111. . . Binary target decision source data

12. . . Skewed coefficient calculation module

13. . . Trust interval calculation module

14. . . Graphic module

Claims (12)

A method for determining a binary target using a confidence interval of a skew coefficient is applicable to a decision result for a user to obtain a binary target, and the method for determining the binary target includes the following steps: (A) providing a set of binary target determination source data (B) using a skew coefficient calculation module, randomly sampling a first predetermined number of samples from the set of binary target data, and calculating a skew coefficient from the sampled samples (C) the skew coefficient calculation module repeatedly performs the step (B) until a second predetermined number of skewness coefficients are obtained; (D) using a confidence interval calculation module from the second predetermined number a skewed coefficient, a set of trust intervals of the skewed coefficients; and (E) graphically presenting the confidence interval of the skewed coefficients to represent the user The result of the determination of the required binary target. The method for determining a binary target of a confidence interval using a skew coefficient according to the first aspect of the patent application, wherein in the steps (B) and (C), the skew coefficient calculation module utilizes a shoe sampling method To obtain the skew coefficient of the second predetermined number. The method for determining a binary target using a confidence interval of a skew coefficient according to the first aspect of the patent application, wherein in the step (B), the skew coefficient calculation module uses a third-order dynamic difference method The sampled samples are calculated and the skew coefficient is calculated. According to the method for determining a binary criterion of a confidence interval using a skew coefficient according to the first aspect of the patent application, in the step (D), the confidence interval of the group is a 95% confidence interval. According to the method for determining the binary target of the confidence interval using the skew coefficient according to the first aspect of the patent application, in the step (E), the determination result of the binary target is normal and abnormal. According to the method for determining the binary target of the confidence interval using the skew coefficient according to the first aspect of the patent application, in the step (E), the determination result of the binary target is satisfactory and unsatisfactory. A binary target determination system using a confidence interval of a skew coefficient is suitable for a user to obtain a binary target determination result, the binary target determination system comprising: a source data database for providing a set of binary targets Determining source data; a skew coefficient calculation module for randomly sampling a first predetermined number of samples from the set of binary target determination source data, and calculating a skewed state from the sampled samples a coefficient, and then repeating the steps of randomly sampling and calculating the skew coefficient until a second predetermined number of skewed coefficients are obtained; a confidence interval calculation module for using the second predetermined number of skew coefficients, Calculating a set of confidence intervals of the skewed coefficients; and a graphical module for graphically presenting the confidence interval of the skewed coefficients to represent the binary target determination required by the user . a binary target determination system using a confidence interval of a skew coefficient according to claim 7 of the patent application scope, wherein the skew coefficient calculation module uses a shoe sampling method to obtain the second predetermined number of skewness coefficients . a binary target determination system using a confidence interval of a skew coefficient according to claim 7 of the patent application scope, wherein the skew coefficient calculation module uses a third-order dynamic difference method from the sampled samples Calculate the skew coefficient. The binary target determination system using the confidence interval of the skew coefficient according to item 7 of the patent application scope, wherein the confidence interval of the group is a 95% confidence interval. According to the method of claim 7, the binary target determination system using the confidence interval of the skew coefficient, wherein the binary target is determined to be normal and abnormal. According to the method for determining the binary target of the confidence interval of the skew coefficient according to Item 7 of the patent application scope, the determination result of the binary target is satisfactory and unsatisfactory.