CN112599119B - Method for establishing and analyzing mobility dysarthria voice library in big data background - Google Patents

Abstract

The invention relates to a method for establishing and analyzing a speech library for motor dysarthria under the background of big data, which includes the following steps: design of pronunciation text; speech recording; parameter analysis of speech files; establishment of a database management system; Data analysis for data technology. This invention aims to study the speech characteristics of patients with motor dysarthria caused by neurological diseases. Relying on the advantages of an open network platform, it can realize measurement covering large-scale groups and the collection of relevant information, and realize the speech of Mandarin, dialects, and healthy people. , patients' voices and other phonetic databases, and on this basis, establish a vocabulary database that meets the condition diagnosis of patients with motor dysarthria.

Description

Establishment and analysis method of speech database for motor dysarthria under the background of big data

Technical field

The invention relates to a method for establishing and analyzing a speech library for motor dysarthria in the context of big data.

Background technique

(1) Current status of research on motor dysarthria:

Dysarthria refers to a group of speech disorders caused by muscle control disorders caused by damage to the central nervous system or peripheral nervous system. Motor dysarthria often manifests as slowing, weakening, imprecision, and uncoordination of speech-related muscle tissue movements. It may also affect breathing, resonance, laryngeal voice control, articulation, and rhythm. Clinically, it is often referred to as dysarthria. . Common causes of motor dysarthria include brain trauma, cerebral palsy, amyotrophic lateral sclerosis, multiple sclerosis, stroke, Parkinson's disease, spinocerebellar ataxia, etc. Dysarthria can be divided into flaccid, spastic, dysregulated, hypokinetic, hyperkinetic and mixed types based on neuroanatomical and speech acoustic characteristics. Among communication disorders related to brain injury, the incidence of dysarthria is as high as 54%. At present, clinical examinations of voice, resonance, rhythm, etc. can reflect the speech acoustic characteristics of dysarthria from both subjective and objective aspects, which is conducive to providing targeted treatment and comprehensively and scientifically elucidating the speech acoustic pathology of dysarthria. mechanism.

There are few reports on the overall incidence of motor dysarthria at home and abroad. Miller et al. studied 125 patients with Parkinson's disease and showed that 69.6% of the patients had lower average speech intelligibility than the normal control group, of which 51.2% patients one standard deviation lower, indicating a higher prevalence of dysarthria in patients with Parkinson's disease. Bogousslavsky et al. screened 1,000 first-time stroke patients and found that up to 46% of patients had speech impairment, of which 12.4% were diagnosed with dysarthria. Hartelius et al. also found that the incidence of dysarthria in patients with multiple sclerosis was 51%. This shows that the incidence of dysarthria is relatively high. There is currently no unified assessment method in China for the assessment of dysarthria, and there is no specific assessment standard for motor dysarthria. Most use the Frenchay Dysarthria Evaluation Method or modified method and the China Rehabilitation Research Center Dysarthria Checklist, which is developed by clinical The physician or rehabilitation physician examines, scores, records, and evaluates the degree and type of dysarthria.

(2) Current status of domestic phonetic database research:

With the development of information technology and computer science, speech technology makes it possible to interact between machine behavior and human natural language. Whether it is speech synthesis, speech recognition or speech recognition research, it must rely on the construction of excellent back-end speech corpora. At present, the development of foreign phonetic libraries is relatively mature, and China's phonetic bank research has also advanced by leaps and bounds in the past ten years. The research and establishment of phonetic banks have been implemented in different language and cultural contexts. However, the construction of a speech library for motor dysarthria is still in a state of research.

Domestic research on the assessment of articulation and speech function mainly focuses on subjective assessment, and only a few researchers distinguish the concepts of articulation and speech. Huang Zhaoming et al. proposed the "Chinese Articulation Ability Test Word List", which contains 50 words. By evaluating the subjects' articulation pronunciation of 50 words, the speech rehabilitation therapist can comprehensively evaluate the subjects' understanding of 21 initial consonants and 4 types of sounds. At the same time, the subject's phoneme contrast ability was assessed through 18 phoneme comparisons and 37 minimum phonetic pairs. Chen Sanding and others evaluated the initial consonants, finals and tones of Mandarin Chinese on 50 deaf children, revealing the development rules of articulation and pronunciation of deaf children who speak Mandarin Chinese, and further proposed the speech rehabilitation of "early, sequential, fault-tolerant and consolidated" Educational principles. Dr. Zhang Jing from East China Normal University studied the main errors made by hearing-impaired children in the articulation of consonants, analyzed the causes, and accordingly proposed a consonant phoneme treatment framework for hearing-impaired children.

(3) Current research status of big data in the medical field:

Currently, the more popular definition of big data is: data that exceeds the capabilities of typical database software tools to capture, store, process, and analyze. Big data is different from traditional data concepts such as ultra-large-scale data and massive data. It has four basic characteristics: large amount, diversity, timeliness, and value. Kayyali B et al. studied the impact of big data on the U.S. medical industry and pointed out that as time goes by, the value of big data to the medical industry will become increasingly significant. At present, big data in the medical field mainly comes from pharmaceutical companies, clinical diagnosis data, patient medical data, health management, and social network data. For example, drug research and development is a relatively intensive process. Even for small and medium-sized enterprises, the data of a drug research and development is more than TB. The data of hospitals are also growing very fast every day. A patient's dual-source CT examination can produce 3,000 images at a time. , approximately 1.5GB of imaging data is generated, and a standard pathological examination image has nearly 5GB of images. Together with patient medical treatment, electronic medical records and other data, it is growing rapidly every day. Research methods based on massive big data analysis have triggered people's thinking about scientific methodology. Research does not require direct contact with the research object, but new research discoveries can be obtained by directly analyzing and mining massive data. This may give rise to a new scientific research model.

The establishment of a speech corpus is a tedious and complicated issue, and there are still issues that need to be improved in the later improvement of the speech corpus, such as making full use of the existing inter-word inflection rules and trying to reflect the actual situation of inflection and softness. For the shortage of corpus, the utilization rate of existing corpus can be improved in the preprocessing step. In view of the above reasons, the speech database should be an open database so that it can be added and modified at any time to improve the database. Since speech conditions are different, the establishment of specific speech corpora will also encounter various difficulties. The issues we discuss here are just a discussion on the establishment of speech corpora. We hope that they can provide information for speech research. Data support plays an important role in better developing language and improving speech corpus.

In addition, the large amount of data is undoubtedly a major advantage of network big data analysis technology, but how to ensure the quality of massive data and how to clean, manage and analyze massive data have also become a major issue in this research. Technical Difficulties. Massive network big data has the characteristics of multi-source heterogeneity, interactivity, timeliness, burstiness and high noise. This leads to the characteristics that although network big data is huge in value, it is also noisy and has low value density. This poses a huge challenge to ensuring data quality in network big data analysis research.

Contents of the invention

The present invention designs a method for establishing and analyzing a speech library for motor dysarthria in the context of big data. The technical problem it solves is that large amounts of data are undoubtedly a major advantage of network big data analysis technology, but how to ensure massive amounts of data? The quality of data and how to clean, manage and analyze massive data have also become a major technical difficulty.

In order to solve the above-mentioned technical problems, the present invention adopts the following solutions:

A method for establishing and analyzing a speech library for motor dysarthria in the context of big data, including the following steps: Step 1. Design of pronunciation text;

Step 2. Voice recording;

Step 3. Annotation of voice files;

Step 4. Analyze the acoustic parameters of the speech file;

Step 5. Establishment of database management system;

Step 6. Data analysis using big data technology.

Preferably, the data analysis of big data technology in step 6 is based on the speech classification mechanism of the Hadoop platform, specifically including the following sub-steps:

Step 61: Collect multiple patient voice files, segment and annotate the speech segments, build a voice database, analyze the extracted acoustic parameters, and obtain effective features for voice classification;

Step 62. Then based on the Hadoop platform, use the Map function to subdivide the big data speech classification problem, use multiple nodes to solve the speech classification problem in parallel and distributed manner, and obtain the corresponding speech classification results;

Step 63: Finally, use the Reduce function to combine the speech classification results of the sub-problems to adapt to the online requirements of big data speech classification.

Preferably, the design of the pronunciation text in step 1 includes the selection of the pronunciation text, and the selection principles of the corpus of the pronunciation text include one or more of the following:

a. The words in the corpus are required to include all phonological phenomena as much as possible, so as to better and more conveniently reflect the phonological characteristics of different patients’ voices;

b. The vocabulary in the corpus is based on commonly used tables for Chinese surveys, so it can be easily compared with Mandarin Chinese;

c. The sentences in the corpus are mainly obtained from conversations with patients based on several related topics, so they are more in line with the real situations faced by speech recognition; "several related topics" include daily life topics or medical history topics, such as asking about the time of the first onset of illness and medical history.

d. The sentences in the corpus are complete in content and semantics, so they can reflect the prosodic information of a sentence as much as possible;

e. The triphones are not classified into categories, which can effectively solve the problem of sparse training data.

Preferably, the design of the pronunciation text in step 1 also includes the preparation of the pronunciation text, and the preparation principles of the pronunciation text include one or more of the following:

a. Single character part: The initials and finals listed in the survey word list and some common words with tones are used as the main recording materials for this phonetic database;

b. Vocabulary part: Based on but not limited to a 4,000-word vocabulary list, relevant words are recorded according to the original conclusions about related phonology, and strive to fully reflect their phonetic characteristics, including phonetic and super-phonetic characteristics, and target some very important For distinctive phonetic phenomena, examples can be added to reflect their characteristics; "Related phonological conclusions and related words" refer to the summary of commonly used sounds based on the sounds used in the same language, the combination rules, and the characteristics of rhythm and intonation. vocabulary.

"Special phonetic phenomena" refer to those that are easy to mispronounce in dialects, such as flat tongue sounds that are difficult to distinguish from raised tongue sounds, f and h being indifferent, etc.

c. Sentence material part: The number of corpus is determined according to the language mastery of different speakers. When selecting, it is necessary to ensure that the scope of the corpus is as wide as possible and to make it have a certain degree of representativeness; "representative" here means that it can A universal statement that embodies the language characteristics of motor dysarthria.

d. Natural dialogue part: Daily life topics, using the form of answering questions and free conversation, recording 20-40 minutes of the speaker's voice material, involving vocabulary that is different from daily spoken language and Mandarin, and requiring the speaker to speak it in dialect.

Preferably, the voice recording in step 2 includes the determination of the speaker, and the selection principle of the speaker is to select a speaker with clear speech and a moderate speaking speed ("moderate speaking speed" refers to a moderate speaking speed, controlled at 120-150 words/ Minutes), a native speaker who is proficient in the local language and willing to actively cooperate with the investigation, must also ensure that the language environment he or she is in is relatively stable, and at the same time must be well-educated; or/and, the voice recording also includes using a voice collector The speech collection is carried out in two ways: one is reading aloud with prompt text, the prompt is Chinese text material, the speaker converts it into his own mother tongue and reads it aloud; the other is natural speech, pronunciation People use prompts to tell folk stories, ethnic living conditions, and humming of local folk songs.

Preferably, the acoustic parameter analysis of the speech file described in step 4 includes speech annotation of the speech library. The basic speech annotation includes the segmentation and alignment of final consonants of each syllable, and the annotation of phonetic tones, including two parts: the first part. It is text annotation. Chinese characters + pinyin are phonetic transliterations. Speech information is recorded in Chinese characters so that it can be used by the recognition system and can also provide material for linguistic research. Text annotation must indicate basic text information and paralinguistic phenomena. Paralinguistic phenomena in basic annotation can be represented by universal paralinguistic symbols; the second part is syllable annotation. Standard Mandarin syllable annotation is used for Mandarin syllable annotation, and syllable annotation is tonal annotation; in tone annotation, 0 means soft, and 1 means Yinping. 2 means Yangping, 3 means rising tone, and 4 means falling tone.

Preferably, the acoustic parameter analysis of the speech file described in step 4 also includes the extraction of acoustic parameters; first, the recorded speech is segmented and the silent segments are eliminated to ensure that the objects of analysis are single words, phrases, Sentences and dialogues; then the starting and ending segments of the speech signal are determined in the speech waveform data, and the speech is marked; finally, the corresponding fundamental frequency and formant acoustic analysis parameter data are obtained based on the autocorrelation algorithm.

Preferably, the establishment of the database management system in step 5 includes the selection of a database, and the sql database management system, which is easier to implement, is selected.

A big data speech classification process method based on the Hadoop platform, including the following steps: using the above establishment method to construct a speech library, based on this speech library, based on the Hadoop platform, using the Map function to subdivide the big data speech classification problem , use multiple nodes to solve the speech classification of sub-problems in parallel and distributed manner, and obtain the corresponding speech classification results; finally, use the Reduce function to combine the speech classification results of the sub-problems to adapt to the online requirements of big data speech classification.

Specific steps are as follows:

(1) Client submits a speech classification task to the Job Tracker of the Hadoop platform, and the Job Tracker copies the speech feature data to the local distributed file processing system;

(2) Initialize the speech classification task, put the task into the task queue, and the Job Tracker allocates the task to the corresponding node, that is, the Task Tracker, based on the processing capabilities of different nodes;

(3) Each Task Tracker uses a support vector machine to fit the relationship between the speech features to be classified and the speech feature library according to the assigned tasks, and obtains the corresponding category of the speech;

(4) Save the corresponding category of the voice as Key/Value to the local file disk;

(5) If the Key/Value of the intermediate results of speech classification are the same, merge them, pass the merged results to Reduce for processing, obtain the results of speech classification, and write the results to the distributed file processing system;

(6) Job Tracker clears the task status, and the user obtains the speech classification results from the distributed file processing system.

The establishment and analysis method of the motor dysarthria speech database in the context of big data has the following beneficial effects:

(1) The present invention aims to study the speech characteristics of patients with motor dysarthria caused by neurological diseases. Relying on the advantages of an open network platform, it can achieve measurement covering large-scale groups and the collection of relevant information, and realize the implementation of Mandarin, dialect, The establishment of speech databases such as healthy people's voices and patients' voices, and on this basis, the establishment of a vocabulary database that meets the condition diagnosis of patients with motor dysarthria.

(2) With the continuous expansion of the speech database, the present invention finally establishes a rich data resource center based on information such as Mandarin, dialects, different medical histories, different conditions, etc., providing a way for patients with neurological diseases to self-diagnose via the Internet, and can also assist Doctors conduct clinical diagnosis and treatment, providing a rich and accurate data platform for quantifying neurological diseases.

(3) Based on the speech database and the Hadoop platform, the present invention uses the Map function to subdivide the big data speech classification problem, uses multi-node parallel and distributed to solve the speech classification of the sub-problems, and obtains the corresponding speech classification results. ; Finally, the Reduce function is used to combine the speech classification results of the sub-problems to adapt to the online requirements of big data speech classification.

Description of drawings

Figure 1: Example of phonetic annotation of "bao" in the embodiment of the present invention.

Figure 2: Formant data of the "bao" voice in the embodiment of the present invention.

Figure 3: The basic framework of the Hadoop platform in the embodiment of the present invention.

Figure 4: The present invention's big data voice classification process based on the Hadoop platform.

Detailed ways

The present invention will be further described below in conjunction with Figures 1 to 4:

The speech library consists of an unvoiced sound library, a voiced sound library, a tone library, a speech synthesis program, and a Chinese-pinyin conversion program.

1. Creation of unvoiced sound library:

According to the characteristics of unvoiced sounds, in order to improve the quality of synthesized speech. The unvoiced sound library is created using the direct sampling method. That is, the unvoiced parts in front of the voiced segments in various pinyin combinations are sampled to form an unvoiced database. Since the unvoiced sounds in a syllable actually account for only a small part, the unvoiced sound library composed of the unvoiced sounds extracted from more than 400 atonal syllables actually occupies a very small storage space.

2. Creation of voiced sound library:

The voiced sounds are synthesized by calling the VTFR corresponding to the voiced sounds by the voiced sound synthesis program. The voiced sound library is actually composed of VTFRs of various voiced sounds. The VTFR extraction program is used to extract the VTFRs of various voiced sounds in sequence. The VTFRs of various voiced sounds and the voiced sound synthesis program are saved in one data package to form a voiced sound library. The actual extracted VTFR is only one curve, and the voiced sound library constructed in this way occupies very little space.

The establishment of the speech corpus of the present invention mainly includes the following four main processes: design of pronunciation text; speech recording; parameter analysis of speech files; establishment of a database management system; and data analysis of big data technology.

1. Design of pronunciation text;

1.1 Selection of pronunciation text:

How to select corpus is the key to corpus construction. In order to ensure the orderly and effective construction of the corpus and the quality of the corpus, the corpus selection principles must first be studied and formulated before corpus construction. The selection principles of this speech corpus include: 1. The words in the corpus are required to contain all phonetic phenomena as much as possible, so that they can better and more conveniently reflect the phonological characteristics of the dialect's speech; 2. The vocabulary in the corpus is based on the common Chinese survey table: Basics, so it can be easily compared with Mandarin Chinese; 3. The sentences in the corpus are mainly selected from spoken corpus! Therefore, it is more in line with the real situation faced by speech recognition; 4. The sentences in the corpus are complete in content and semantics, so they can reflect the prosodic information of a sentence as much as possible; 5. We do not classify triphones selection, which can effectively solve the problem of sparse training data.

1.2 Preparation of pronunciation text:

The preparation of pronunciation text is one of the key steps in establishing a speech database. When we determine the pronunciation material, we follow the pronunciation text selection principle and include five parts: First, the single-character part. The initials and finals listed in the survey word list and some common words with tones are used as the main recording materials for this phonetic database; the second is the vocabulary part. Based on but not limited to a 4,000-word vocabulary list, relevant words are recorded according to the original conclusions about related phonology, striving to fully reflect their phonetic characteristics, including phonetic and super-phonetic characteristics, and to target some very distinctive phonetic phenomena. , examples can be added to reflect its characteristics; the third is the sentence material part. The number of corpus is determined according to the language mastery of different speakers. When selecting, it is necessary to ensure that the scope of the corpus is as wide as possible and to make it have a certain representativeness; The fourth is the natural dialogue part, which takes the topic of daily life and adopts the form of answering questions and free conversation. The speaker’s voice material is recorded for about half an hour. It involves words that are different in daily spoken language and Mandarin, and the speaker is required to speak them in dialect.

2. Voice recording;

2.1 Determination of the speaker:

The principle of selecting speakers is to select native speakers who are clear in speech, speak at a moderate speed, are proficient in the local language, and are willing to actively cooperate with the investigation. They must also ensure that their language environment is relatively stable and they must have a certain level of education.

2.2 Voice collection:

The way you speak during the recording process directly determines the purpose of the voice library. Due to the particularity of the collected corpus, two methods are used according to different research purposes: one is reading aloud with text prompts, and the prompts are Chinese text materials! The speaker converts it into his native language and reads it aloud; the other is natural speech, and the speaker can use prompts to tell folk stories, national living conditions, and humming of local folk songs.

3. Parameter analysis of voice files:

After recording the pronunciation text, the speech data needs to be analyzed and processed to obtain different characteristics of the speech signal. This is the key to speech corpus design and the necessary foundation for later speech processing. The present invention focuses on studying speech information, so it is necessary to label the basic attributes of the speech signal waveform and extract relevant acoustic parameters at the same time.

3.1 Information annotation of speech database:

Speech annotation uses Praat software and refers to the Chinese speech segment annotation system SAMPA-C for hierarchical annotation. The annotation of the speech library includes two parts: text annotation and tonal syllable annotation. Here, the phonetic "bao" is taken as an example, as shown in Figure 1.

The first part is text annotation. Chinese characters + pinyin are phonetic transliterations. The speech information is recorded in Chinese characters so that it can be used by the recognition system and can also provide material for linguistic research. Text annotations must indicate basic text information and paralinguistic phenomena. Paralinguistic phenomena in basic annotations can be represented by universal paralinguistic symbols.

The second part is syllable annotation. Standard Mandarin syllable annotation is used for Mandarin syllable annotation, and the syllable annotation is tonal annotation. In the tone annotation, 0 means soft tone, 1 means yin level, 2 means yang level, 3 means up tone, and 4 means down tone.

3.2 Extraction of acoustic parameters:

For the recorded speech signal, it is necessary to extract the acoustic parameters of each segment. In actual operation, the recorded speech is first segmented and the silent segments are eliminated to ensure that the analyzed objects are single words; then In the speech waveform data, the starting and ending segments of the speech signal are determined, and the range of finals is marked; finally, the corresponding fundamental frequency and formant data are obtained based on the autocorrelation algorithm, taking the speech "bao" as an example, as shown in Figure 2.

4. Establishment of database management system:

4.1 Selection of database

Regarding the choice of database, a large amount of voice waveform data needs to be stored in the speech database, which is characterized by a large amount of data, variable length, and low requirements for transaction processing and recovery, security, and network support. Therefore, we can choose a SQL database management system that is easier to implement.

4.2 Establishment of database management system

The establishment of the database management system in the speech corpus requires the storage of four types of materials: first, the speaker's attribute materials, such as the speaker's age, gender, education status, mastery of Chinese, and the person's use of the mother tongue, etc.; the second is the pronunciation text material. Enter and store the speaker's pronunciation materials and their corresponding dialect pronunciation and Mandarin International Phonetic Alphabet and other text materials; the third is the actual speech data material, which is mainly used to save the original parameters of the recorded speech waveform graphics; the fourth is the acoustic analysis parameter data, that is Save the acoustic parameters extracted from the processed speech waveform.

5. Data analysis using big data technology

Big data is a collection of data that is so large that its acquisition, storage, management, and analysis greatly exceed the capabilities of traditional database software tools. It has massive data scale, rapid data flow, diverse data types, and low value density. Big features. The strategic significance of big data technology lies not in mastering huge data information, but in professional processing of these meaningful data. In other words, if big data is compared to an industry, then the key to making this industry profitable is to improve the "processing capabilities" of data and achieve the "value-added" of data through "processing". In the construction of the lexicon, the important value of using big data technology is to achieve the purpose of evaluating the quality of the phonetic elements in the lexicon through targeted analysis and research of the data, thereby making the construction of the lexicon more complete.

The vocabulary library will be shared through the network platform to facilitate testing of different groups of people. At the same time, more data samples will be obtained to enrich the speech library. In the future, more targeted movement articulation can be established based on different regions and different dialects. The dictionary of patients with disorders provides richer and more reliable data samples for subsequent automatic identification of disease classification and grading.

As shown in Figure 3, a speech classification mechanism based on the Hadoop platform is proposed. First, a large number of images are collected, an image database is constructed, and effective features for image classification are extracted. Then based on the Hadoop platform, the Map function is used to solve the big data speech classification problem. Subdivided into sub-problems, multiple nodes are used to solve the speech classification problem in parallel and distributed manner, and the corresponding speech classification results are obtained. Finally, the Reduce function is used to combine the speech classification results of the sub-problems to adapt to the online requirements of big data speech classification.

As shown in Figure 4, the specific steps of the big data speech classification process based on the Hadoop platform are as follows:

(1) Client submits a speech classification task to the Job Tracker of the Hadoop platform, and the Job Tracker copies the speech feature data to the local distributed file processing system;

(2) Initialize the speech classification task, put the task into the task queue, and the Job Tracker allocates the task to the corresponding node, that is, the Task Tracker, based on the processing capabilities of different nodes;

(3) Each Task Tracker uses a support vector machine to fit the relationship between the speech features to be classified and the speech feature library according to the assigned tasks, and obtains the corresponding category of the speech;

(4) Save the corresponding category of the voice as Key/Value to the local file disk;

(5) If the Key/Value of the intermediate results of speech classification are the same, merge them, pass the merged results to Reduce for processing, obtain the results of speech classification, and write the results to the distributed file processing system;

(6) Job Tracker clears the task status, and the user obtains the speech classification results from the distributed file processing system.

The present invention has been exemplarily described above in conjunction with the accompanying drawings. It is obvious that the implementation of the present invention is not limited by the above-mentioned manner, as long as various improvements of the method, concept and technical solution of the present invention are adopted, or the present invention is implemented without improvement. If the concepts and technical solutions are directly applied to other situations, they are all within the protection scope of the present invention.

Claims (4)

1. A method for establishing and analyzing a mobility dysarthria voice library under a big data background comprises the following steps:

step 1, designing a pronunciation text;

step 2, recording voice;

step 3, labeling the voice file;

step 4, analyzing acoustic parameters of the voice file;

in the step 4, the acoustic parameter analysis of the voice file includes voice labeling of a voice library, the basic voice labeling includes initial and final segmentation and alignment of each syllable, and labeling of initial and final tones, including two parts:

the first part is character labeling, chinese characters and pinyins are word-sound transcription, voice information is recorded by Chinese characters so as to be provided for a recognition system to use, and materials can also be provided for linguistic study; the character label must mark basic character information and secondary linguistic phenomena, and the secondary linguistic phenomena in the basic label can be represented by general secondary linguistic symbols;

the second part is syllable labeling, standard Mandarin syllable labeling is adopted for the Mandarin syllable labeling, and the syllable labeling is tone labeling; in the tone mark, 0 represents light sound, 1 represents shade level, 2 represents sun level, 3 represents rising sound, and 4 represents falling sound;

the acoustic parameter analysis of the voice file in the step 4 further comprises extraction of acoustic parameters;

firstly, cutting and eliminating a mute section of recorded voice to ensure that an analyzed object is a single word, phrase, sentence and dialogue; then, judging the start and stop sections of the voice signal in the voice waveform data, and marking the voice; finally, corresponding fundamental frequency and formant acoustic analysis parameter data are obtained according to an autocorrelation algorithm;

step 5, establishing a database management system;

the establishment of the database management system in the step 5 comprises the selection of a database, and the selection of an sql database management system which is easy to realize is adopted; in the step 5, four materials are needed to be stored in the establishment of the database management system, namely, attribute materials of a speaker; secondly, inputting and storing the pronunciation text material of the patient and corresponding pronunciation thereof, the international phonetic symbols of Mandarin and other text materials; third, the actual voice data material is used for storing the original parameters of the recorded voice waveform graph; fourthly, acoustic analysis parameter data, namely, saving acoustic parameters extracted from the processed voice waveform;

step 6, data analysis of big data technology;

the data analysis of the big data technology in the step 6 is based on a voice classification mechanism of a Hadoop platform, and specifically comprises the following sub-steps:

step 61, collecting a plurality of patient voice files, segmenting and labeling voice segments, constructing a voice database, analyzing extracted acoustic parameters, and obtaining effective characteristics of voice classification;

step 62, subdividing the big data voice classification problem by using a Map function based on the Hadoop platform, and carrying out voice classification solving on the sub-problem in a parallel and distributed manner by using multiple nodes to obtain a corresponding voice classification result;

and 63, finally, combining the voice classification results of the sub-problems by using a Reduce function so as to adapt to the online requirements of large data voice classification.

2. The method for establishing and analyzing the dyskinetic voice library under the big data background according to claim 1, wherein the method comprises the following steps of:

the design of the pronunciation text in the step 1 comprises the selection of the pronunciation text, and the selection principle of the corpus of pronunciation text comprises one or more of the following:

a. the single word in the corpus is required to contain all the initial and final phenomena as much as possible, so that the sound system characteristics of the voices of different patients can be reflected better and more conveniently;

b. the vocabulary in the corpus is based on the Chinese investigation common table, so that the vocabulary can be conveniently compared with Mandarin Chinese;

c. sentences in the corpus are mainly obtained by dialogue with a patient according to a plurality of related topics, so that the sentences more conform to the real situation of speech recognition;

d. sentences in the corpus are complete in content and semantics, so that prosodic information of one sentence can be reflected as far as possible;

e. the triphones are not classified, so that the problem of sparse training data can be effectively solved.

3. The method for establishing and analyzing the dyskinetic voice library in the big data background according to claim 2, wherein the method comprises the following steps:

the design of the pronunciation text in the step 1 further comprises the preparation of pronunciation text, and the preparation principle of pronunciation text comprises one or more of the following:

a. single word part: taking common words of initials and finals listed in the investigation word list as the corpus used for the main recording of the voice library;

b. vocabulary part: based on at least one four-thousand word vocabulary, recording related words according to the original conclusion about related sound system, aiming at comprehensively reflecting the voice characteristics, including tone quality and super-tone quality characteristics, and adding example words to reflect the characteristics of some very characteristic voice phenomena;

c. statement materials section: determining the corpus quantity according to the language mastering degree of different pronounciators, wherein the corpus is selected to have a certain representativeness while ensuring the scope of the corpus as wide as possible;

d. natural dialogue portion: the daily life is the question, the form of answering questions and free talking is adopted, the voice material of the speaker for 20-40 minutes is recorded, the words which are different from the common speaking in the daily spoken language are related, and the speaker is required to speak in the dialect.

4. The method for establishing and analyzing the mobility dysarthria voice library in the big data background according to claim 3, wherein the method comprises the following steps:

the voice recording in the step 2 comprises the determination of a speaker, and the selection principle of the speaker is that the speaker with clear mouth and teeth, moderate speed of speech, proficiency in using local language and willingness to actively cooperate with investigation is selected, the language environment of the speaker is ensured to be stable, and the speaker has cultural level;

or/and, the voice recording further comprises voice collection through a voice collector, and the voice collection adopts two modes: one is reading with prompt text, the prompt is a Chinese character material, and a speaker converts the Chinese character material into own native language and reads the native language; the other is natural speech, and the speaker uses cues to tell the folk stories, the state of national life, and humming of local folk songs.