Method and system for predicting cognitive competence of old people based on voice task

Abstract

The embodiment of the invention provides a method and a system for predicting cognitive competence of old people based on voice tasks. The method is applied to the technical field of data processing, and comprises the following steps: collecting voice data and cognitive ability data of old people, performing data perception, and constructing to obtain a voice cognitive ability feature sample of the old people; an integrated robust depth random configuration network model is constructed in a node increment mode, the network model is composed of a plurality of base learners, and each learner comprises a plurality of hidden layers; and performing integrated robust prediction on the speech cognitive ability feature sample according to the network model to obtain a network model integrated prediction result. In this way, the cognitive assessment scale and the voice data are adopted, the integrated robust depth random configuration network is constructed to achieve cognitive prediction of the old people, and the technical problems that in the prior art, cognition of patients cannot be effectively and accurately diagnosed, and the prediction effect is poor due to the fact that the number of influence factors of the voice data of the patients is large are solved.

Description

Technical Field

[0001] The present disclosure relates to the field of data processing technology, and in particular to a method and system for predicting cognitive abilities of the elderly based on speech tasks. Background Art

[0002] Alzheimer's Disease (AD) is a gradually deteriorating neurological disease characterized by progressive cognitive dysfunction. It causes the death of neurons and the breakdown of connections in the brain, which affects memory, thinking, language, and behavior. Patients may lose their memory, get lost, fail to recognize their loved ones, and experience decreased language skills, which ultimately lead to severe cognitive impairment and the loss of the ability to live independently.

[0003] Currently, although there are many methods for evaluating and diagnosing Alzheimer's disease, they are all affected by factors such as the subject's cultural level and emotional state, and are costly, complex to operate, and not suitable for routine screening. In addition, some methods are somewhat invasive and cannot adapt to today's Alzheimer's disease diagnostic practice. With the development of artificial intelligence technology, early warning of Alzheimer's disease based on speech analysis has the advantages of non-invasiveness, convenience, low cost, and objectivity. However, it lacks the prediction of cognitive ability scores.

[0004] Therefore, it is urgently needed to provide a method and a corresponding system for diagnosing and predicting Alzheimer's disease based on voice data and cognitive data of the elderly. Summary of the invention

[0005] The present invention discloses a method and system for predicting cognitive ability of the elderly based on speech tasks. By adopting a cognitive assessment scale and speech data and constructing an integrated robust deep random configuration network, cognitive prediction of the elderly is achieved, which solves the technical problems that the prior art cannot effectively and accurately diagnose patient cognition and has poor prediction results due to multiple factors affecting patient speech data.

[0006] According to a first aspect of the present disclosure, a method for predicting cognitive ability of the elderly based on speech tasks is provided, comprising the following steps: Collect the voice data and cognitive ability data of the elderly and perform data perception to construct a sample of the characteristics of the elderly's voice cognitive ability; An integrated robust deep random configuration network model is constructed by using a node increment method, wherein the network model is composed of a plurality of base learners, and the base learners include a plurality of hidden layers; A predictor is constructed based on the hidden layer, a residual probability density function is constructed based on the speech recognition ability feature sample using a kernel density estimation method, a sample residual matrix of the predictor is calculated, and a weighting matrix of the base learner is obtained according to the sample residual matrix; A weighted negative correlation learning method is adopted and integrated robust prediction is performed based on the weighted matrix to obtain a network model output result; The weight distribution of the network model output result is updated, and the network model integrated prediction result is calculated.

[0007] According to the above aspects and any possible implementation, an implementation is further provided, wherein the process of collecting the voice data and cognitive ability data of the elderly and performing data perception to construct the voice cognitive ability feature sample of the elderly is: The cognitive status of the elderly was assessed using a cognitive status measurement table for the elderly with different cognitive abilities, and the cognitive ability data of the elderly were collected; The recording responses of the elderly to the 100-7, 101-7 and 102-7 computing task voice databases are collected to obtain the elderly voice data; A speech cognitive ability feature sample is constructed based on the cognitive ability data and the speech data.

[0008] According to the above aspects and any possible implementation, there is further provided an implementation, wherein the elderly cognitive status measurement scale includes the Mini-Mental State Examination Scale, the Montreal Cognitive Assessment Scale and the Wechsler Adult Intelligence Scale; The recorded answer results include the accuracy of answering questions, average reaction time, reaction time for first answering questions, number of questions answered within 20 seconds, and number of questions answered correctly within 20 seconds.

[0009] According to the above aspects and any possible implementation, an implementation is further provided, wherein the process of using the weighted negative correlation learning method and performing integrated robust prediction based on the weighted matrix to obtain the prediction result of the network model is: Constructing a weighted output matrix of the robust deep random configuration network and a weighted cognitive ability of the robust deep random configuration network according to the weighted output matrix of the base learner, and constructing a loss function based on the weighted output matrix and the weighted cognitive ability; Define a regularized objective function, use the output of the loss function as the input of the objective function, and calculate the output weight of the network model after weighted negative correlation learning; The integrated prediction results of each network model are obtained based on the output weight calculation of the network model.

[0010] According to the above aspects and any possible implementation, an implementation is further provided, wherein the process of updating the weight distribution of the network model output result and calculating the network model integrated prediction result is: Updating the distribution weights of the samples, and calculating the weighted loss of the model output results according to the distribution weights of the samples; The robust deep random configuration network weights are calculated according to the weighted loss, and the network model integrated prediction results are calculated based on the robust deep random configuration network weights and the network model output results.

[0011] According to the above aspects and any possible implementation, a further implementation is provided, in the robust deep random configuration network model, if the model does not reach the set number of hidden layers and the number of hidden layer nodes corresponding to it or does not meet the preset error , then continue to add hidden layer nodes until the number of hidden layer nodes of the model is met.

[0012] According to the above aspects and any possible implementation, an implementation is further provided, wherein the base learner includes a loss function and a regularized objective function, and the loss function is specifically: ; in, Represents a robust deep random configuration network The weighted output matrix of Represents a robust deep random configuration network The weighted cognitive ability of represents the original weights of each robust deep random configuration network; represents the regularization parameter, , B is the output weight of the network model; The regularization objective function is specifically: .

[0013] According to the above aspects and any possible implementation, an implementation is further provided, in which the accuracy of the network model ensemble prediction result is analyzed by using the mean absolute error and the root mean square error, specifically: ; ; in, RMSE is the root mean square error, MAE is the mean absolute error, To predict the results, For the real result, N The amount of data.

[0014] According to a second aspect of the present disclosure, there is provided a system for predicting cognitive ability of the elderly based on speech tasks, comprising: a module for acquiring data of speech cognitive ability of the elderly, a module for building a robust deep random configuration network, and a module for predicting cognitive ability of the elderly; The elderly speech cognitive ability data acquisition module is used to collect the elderly speech data and cognitive ability data and perform data perception to construct the elderly speech cognitive ability feature samples; The robust deep random configuration network construction module is used to construct an integrated robust deep random configuration network model in a node increment manner, wherein the network model is composed of a plurality of base learners, and the learner includes a plurality of hidden layers; The elderly cognitive ability prediction module is used to perform integrated robust prediction on the speech cognitive ability feature samples according to the network model to obtain the network model integrated prediction results.

[0015] According to the above aspects and any possible implementation, there is further provided an implementation, wherein the robust deep random configuration network construction module includes a weighted matrix calculation module and a robust prediction module; The weighting matrix calculation module is used to construct a predictor based on the hidden layer, adopt a kernel density estimation method and construct a residual probability density function based on the speech cognition ability feature sample, calculate and obtain a sample residual matrix of the predictor, and obtain a weighting matrix of the base learner according to the sample residual matrix; The robust prediction module is used to adopt a weighted negative correlation learning method and perform integrated robust prediction based on the weighted matrix to obtain a network model output result.

[0016] Compared with the prior art, the present invention has the following technical effects: (1) The present invention relies on the Mini-Mental State Examination Scale for the Elderly with different cognitive abilities, the Montreal Cognitive Assessment Scale, the Wechsler Adult Intelligence Scale, etc., which can evaluate the patient's cognitive function, including attention, memory, language ability, executive function and other fields, and has the advantages of being non-invasive, low cost and easy to operate; (2) The integrated robust deep random configuration network of the present invention is composed of multiple base learners, and the base learners are based on The loss function and The objective function of the regularization is composed of calculating the sample weighting function under each base learner based on the kernel density estimation (KDE), and using the weighted negative correlation learning (WNCL) to achieve integrated robust prediction, thereby effectively improving the accuracy and robustness of the model; (3) The present invention can collect data through different terminals such as personal computers, home televisions, mobile phones, and tablets, use cloud servers to analyze and process the data, and provide cognitive ability prediction results in real time, achieving significant improvements and breakthroughs in implementation methods, perception quality, friendly interaction, and health management.

[0017] It should be understood that the contents described in the summary of the invention are not intended to limit the key or important features of the embodiments of the present disclosure, nor are they intended to limit the scope of the present disclosure. Other features of the present disclosure will become easily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] The above and other features, advantages and aspects of the embodiments of the present disclosure will become more apparent with reference to the following detailed description in conjunction with the accompanying drawings. The accompanying drawings are used to better understand the present solution and do not constitute a limitation of the present disclosure. In the accompanying drawings, the same or similar reference numerals represent the same or similar elements, among which: Figure 1 A schematic diagram of a method for predicting cognitive ability of the elderly based on speech tasks according to an embodiment of the present disclosure is shown; Figure 2 A schematic diagram of a robust deep random configuration network model integrated with a method for predicting cognitive ability of the elderly based on speech tasks according to an embodiment of the present disclosure is shown; Figure 3 A schematic diagram of the structure of a system for predicting cognitive ability of the elderly based on speech tasks according to an embodiment of the present disclosure is shown. DETAILED DESCRIPTION

[0019] In order to make the purpose, technical solution and advantages of the embodiments of the present disclosure clearer, the technical solution in the embodiments of the present disclosure will be clearly and completely described below in conjunction with the drawings in the embodiments of the present disclosure. Obviously, the described embodiments are part of the embodiments of the present disclosure, not all of the embodiments. Based on the embodiments in the present disclosure, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present disclosure.

[0020] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the present invention is further described in detail below with reference to the accompanying drawings and specific embodiments.

[0021] The early stage of Alzheimer's Disease (AD) is mild cognitive impairment (MCI). At this stage, patients have normal daily living abilities, but there is a progressive decline in cognitive function. From a treatment perspective, AD is irreversible and difficult to treat, but if patients can be treated at the MCI stage, the onset of dementia can be effectively delayed.

[0022] The random configuration network is a lightweight neural network model based on random weights, which has the advantages of fast modeling speed and good generalization performance. The objective function of the loss is easily affected by noise and outliers. In order to improve the robustness of the model, the robustness of the model has been affected by problems such as acquisition equipment and human factors for complex and diverse speech data. Therefore, the present invention proposes an integrated robust deep random configuration network for predicting cognitive ability of the elderly. The integrated robust deep random configuration network consists of multiple base learners, and the base learners are based on The loss function and The regularized objective function is composed of kernel density estimation (KDE) to calculate the sample weighting function under each base learner, and weighted negative correlation learning (WNCL) is used to achieve integrated robust prediction.

[0023] Reference Figure 1 As shown, this embodiment provides a method for predicting cognitive ability of the elderly based on speech tasks, comprising the following steps: S101. Collect the speech data and cognitive ability data of the elderly and perform data perception to construct a sample of the speech cognitive ability characteristics of the elderly.

[0024] In this embodiment, the system is supported by data such as the Mini-Mental State Examination Scale for Elderly People with different cognitive abilities, the Montreal Cognitive Assessment Scale, the Wechsler Adult Intelligence Scale, and voice data, wherein the voice data is a recording of calculation tasks such as 100-7, 101-7, and 102-7 performed continuously by the patient.

[0025] Specifically, the cognitive impairment caused by AD will affect the ability to express language, which is then reflected in the process and content of language expression. Therefore, by collecting the voice of the user performing a preset descriptive task (such as collecting voice through the microphone of an electronic device) to obtain voice information, the algorithm based on voice analysis recognizes the user's language expression ability and then evaluates the degree of cognitive impairment of the user, thereby detecting AD and MCI based on voice information.

[0026] Among them, the preset description task may include describing the content of the preset image and describing the names of different target objects of the target type within a preset time. It can be understood that impaired cognitive function will affect the accuracy of the description of the preset image. For the content in the image, the more severe the cognitive function is impaired, the worse the accuracy of the description is. When performing the description task, the display screen of the electronic device can display the preset image. The accuracy of the description of the preset image by a normal user and a user with impaired cognitive function is obviously different; similarly, impaired cognitive function will also affect the fluency of the description. For example, the names of different target objects of the target type described within a preset time can be to say as many animal names as possible within a preset time (such as 30 seconds, 1 minute, 2 minutes, etc.). Therefore, by collecting the voice information of the user performing the preset description task, it can be used to detect the degree of cognitive impairment, thereby detecting the probability of the user suffering from AD and MCI.

[0027] The specific process of obtaining the recorded data is as follows: log in to the system through a mobile phone or tablet, and enter the 100-7, 101-7, and 102-7 task answering interfaces in turn (the diversity of answers can be used to objectively evaluate cognitive ability), and answer the calculation results of 100-7, 101-7, 102-7, etc. through the recording function. Taking 100-7 as an example, the correct answer sequence is 93, 86, 79, 72, and so on. Based on the user's accuracy in answering questions, the average reaction time for answering questions, the reaction time for the first answer, the number of questions answered within 20 seconds, and the number of questions answered correctly within 20 seconds, the voice data of the elderly can be finally obtained.

[0028] In this embodiment, the training data is set as ,in, Represents the input data, , d Represents the feature dimension; Represents cognitive ability, , Representing the qualities of different scales, the cognitive results of scales in multiple dimensions can be used to make multi-output predictions; , N Represents the number of training samples.

[0029] S102. Constructing an integrated robust deep random configuration network model by using a node increment method, wherein the network model is composed of a plurality of base learners, and the learner includes a plurality of hidden layers.

[0030] like Figure 2 As shown, this embodiment adopts a robust deep random configuration network and adopts a node increment method for modeling, assuming that n hidden layer, n Layer The nodes have been configured, and the current network output can be expressed as , the training error is defined as : If the current model does not reach the set number of hidden layers And the corresponding number of hidden layer nodes , or the preset error is not met , continue to increase hidden layer nodes, Randomly generate multiple hidden layer nodes Weight and bias , select The maximum result is taken as the final parameter, where Calculation is as follows: (1) in, Represents the hidden layer nodes The output of and bias , calculated by the activation function and the output of the previous layer of network; ; ; .

[0031] S103. Performing integrated robust prediction on the speech cognition ability feature samples according to the network model to obtain an integrated prediction result of the network model.

[0032] In order to improve the generalization performance of the model and the ability to handle outliers, formula (2) is used The loss function of the paradigm is: (2) in, represents the basis model weight, Represents cognitive ability, Representative base model n The output of the hidden layer, represents the base model regularization parameter.

[0033] Secondly, to further improve the model performance, this embodiment designs a weighted negative correlation learning strategy based on kernel density estimation for model integration, where the number of predictors is K , each predictor is represented by k , k =1, 2, …, K .

[0034] In addition, compared with the residuals of normal samples, the residuals of vertical outliers are always far away from the center of the overall residuals and have low density. Therefore, the probability density of the sample residuals can be calculated by kernel density estimation (KDE) to determine the weight matrix of each base learner. , so that normal data gets a higher weight and noise data gets a smaller weight.

[0035] The residual probability density function based on kernel density estimation (KDE) is as follows: (3) in, Indicates the estimated window width; Representative samples The residual of represents the kernel function, usually a Gaussian kernel function, N is the number of samples.

[0036] Finally, the residual matrix of each sample of predictor k can be obtained: (4) in, is the sample weight matrix of the predictor, is the sample residual.

[0037] Then, to achieve ensemble learning, the loss function of k predictors is Minimum is the goal, and the objective function (5) is defined, where Represents the output weight of the integrated model after weighted negative correlation learning.

[0038] (5) Among them, the loss function It can be defined as formula (6): (6) in, Represents a robust deep random configuration network The weighted output matrix of Represents a robust deep random configuration network weighted cognitive ability; represents the original weights of each robust deep random configuration network; represents the regularization parameter; .

[0039] According to the solution of formula (6), we can get , then the results of each base model can be calculated , in order to perform model integration, according to formula (7) Weighted loss of regression results , update the distribution weight of each sample according to formula (8): (7) (8) (9) in, Represents the sample distribution weight, the initial value is , Representation predictor For samples The predicted value of Representation sample The actual cognitive ability value, represents the threshold function, which is used to update the distribution weight; represents the weight of the robust deep random configuration network k; According to formula (10), the final integrated model output result can be obtained: (10) like Figure 3 As shown, this embodiment also provides a system for predicting cognitive ability of the elderly based on speech tasks, including: an elderly speech cognitive ability data acquisition module 1, a robust deep random configuration network construction module 2 and an elderly cognitive ability prediction module 3; The elderly speech recognition ability data acquisition module 1 is used to collect the elderly speech data and recognition ability data and perform data perception to construct the elderly speech recognition ability feature samples.

[0040] The robust deep random configuration network construction module 2 is used to construct an integrated robust deep random configuration network model using a node increment method, wherein the network model is composed of a number of base learners, and the learner contains a number of hidden layers.

[0041] The robust deep random configuration network construction module 2 includes a weighted matrix calculation module 21 and a robust prediction module 22; The weighting matrix calculation module 21 is used to construct a predictor based on the hidden layer, use the kernel density estimation method and construct a residual probability density function based on the speech cognition ability feature sample, calculate the sample residual matrix of the predictor, and obtain the weighting matrix of the base learner according to the sample residual matrix; The robust prediction module 22 is used to adopt a weighted negative correlation learning method and perform integrated robust prediction based on a weighted matrix to obtain a network model output result.

[0042] The elderly cognitive ability prediction module 3 is used to perform integrated robust prediction on the speech cognitive ability feature samples according to the network model to obtain the network model integrated prediction results.

[0043] Specifically, this embodiment collects data of elderly people with different cognitive abilities based on a tertiary-level A-class hospital, records Montreal cognitive assessment (MoCA) scale data, and collects voice data of 100-7 tasks based on voice. By performing feature extraction and using the prediction results obtained by the method of the present invention, the cognitive ability prediction error is: mean absolute error (MAE) 3.06, root mean square error (RMSE) 3.92, and a good prediction result is achieved.

[0044] During the implementation of the case, users use mobile phones and other platforms to input syllables according to the guidance. After the input syllables are preprocessed, the predictor module and the multi-syllable fusion diagnosis module can obtain the probability of Alzheimer's disease under single syllables and the probability of Alzheimer's disease under polysyllables, and can provide guidance such as precautions based on the probabilities.

[0045] It should be noted that, for the aforementioned method embodiments, for the sake of simplicity, they are all described as a series of action combinations, but those skilled in the art should be aware that the present disclosure is not limited by the order of the actions described, because according to the present disclosure, certain steps can be performed in other orders or simultaneously. Secondly, those skilled in the art should also be aware that the embodiments described in the specification are all optional embodiments, and the actions and modules involved are not necessarily required by the present disclosure.

[0046] It should be understood that the various forms of processes shown above can be used to reorder, add or delete steps. For example, the steps recorded in this disclosure can be executed in parallel, sequentially or in different orders, as long as the desired results of the technical solutions disclosed in this disclosure can be achieved, and this document does not limit this.

[0047] The above specific implementations do not constitute a limitation on the protection scope of the present disclosure. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions can be made according to design requirements and other factors. Any modification, equivalent substitution and improvement made within the spirit and principle of the present disclosure shall be included in the protection scope of the present disclosure.

Claims

1. A method for predicting cognitive ability of the elderly based on speech tasks, characterized in that: The following steps are involved: Collect the voice data and cognitive ability data of the elderly and perform data perception to construct a sample of the characteristics of the elderly's voice cognitive ability; An integrated robust deep random configuration network model is constructed by using a node increment method, wherein the network model is composed of a plurality of base learners, and the base learners include a plurality of hidden layers; A predictor is constructed based on the hidden layer, a residual probability density function is constructed based on the speech recognition ability feature sample using a kernel density estimation method, a sample residual matrix of the predictor is calculated, and a weighting matrix of the base learner is obtained according to the sample residual matrix; A weighted negative correlation learning method is adopted and integrated robust prediction is performed based on the weighted matrix to obtain a network model output result; The weight distribution of the network model output result is updated, and the network model integrated prediction result is calculated.

2. The method for predicting cognitive ability of the elderly based on speech tasks according to claim 1 is characterized in that: The process of collecting the voice data and cognitive ability data of the elderly and performing data perception to construct the voice cognitive ability feature samples of the elderly is as follows: The cognitive status of the elderly was assessed using a cognitive status measurement table for the elderly with different cognitive abilities, and the cognitive ability data of the elderly were collected; The recording responses of the elderly to the 100-7, 101-7 and 102-7 computing task voice databases are collected to obtain the elderly voice data; A speech cognitive ability feature sample is constructed based on the cognitive ability data and the speech data.

3. The method for predicting cognitive ability of the elderly based on speech tasks according to claim 2 is characterized in that: The cognitive status measurement scales for the elderly include the Mini-Mental State Examination, the Montreal Cognitive Assessment, and the Wechsler Adult Intelligence Scale; The recorded answer results include the accuracy of answering questions, average reaction time, reaction time for first answering questions, number of questions answered within 20 seconds, and number of questions answered correctly within 20 seconds.

4. The method for predicting cognitive ability of the elderly based on speech tasks according to claim 1 is characterized in that: The process of using the weighted negative correlation learning method and performing integrated robust prediction based on the weighted matrix to obtain the prediction result of the network model is as follows: Constructing a weighted output matrix of the robust deep random configuration network and a weighted cognitive ability of the robust deep random configuration network according to the weighted output matrix of the base learner, and constructing a loss function based on the weighted output matrix and the weighted cognitive ability; Define a regularized objective function, use the output of the loss function as the input of the objective function, and calculate the output weight of the network model after weighted negative correlation learning; The integrated prediction results of each network model are obtained based on the output weight calculation of the network model.

5. The method for predicting cognitive ability of the elderly based on speech tasks according to claim 1 is characterized in that: The process of updating the weight distribution of the network model output result and calculating the network model integrated prediction result is as follows: Updating the distribution weights of the samples, and calculating the weighted loss of the model output results according to the distribution weights of the samples; The robust deep random configuration network weights are calculated according to the weighted loss, and the network model integrated prediction results are calculated based on the robust deep random configuration network weights and the network model output results.

6. The method for predicting cognitive ability of the elderly based on speech tasks according to claim 1 is characterized in that: In the robust deep random configuration network model, if the model does not reach the set number of hidden layers and the corresponding number of hidden layer nodes or does not meet the preset error , then continue to add hidden layer nodes until the number of hidden layer nodes required by the model is met.

7. The method for predicting cognitive ability of the elderly based on speech tasks according to claim 4 is characterized in that: The base learner includes a loss function and a regularized objective function, and the loss function is specifically: ； in, Represents a robust deep random configuration network The weighted output matrix of Represents a robust deep random configuration network The weighted cognitive ability of represents the original weights of each robust deep random configuration network; represents the regularization parameter, , B is the output weight of the network model; The regularization objective function is specifically: 。 8. The method for predicting cognitive ability of the elderly based on speech tasks according to claim 1 is characterized in that: The mean absolute error and root mean square error are used to analyze the accuracy of the network model ensemble prediction results, specifically: ； ； Among them, RMSE is the root mean square error, MAE is the mean absolute error, To predict the results, For the real result, N The amount of data.

9. A system for predicting cognitive ability of the elderly based on speech tasks, implemented by the method for predicting cognitive ability of the elderly based on speech tasks as described in any one of claims 1 to 8, characterized in that: include: Elderly people's speech cognition ability data acquisition module (1), robust deep random configuration network construction module (2) and elderly people's cognitive ability prediction module (3); The elderly speech cognitive ability data acquisition module (1) is used to collect the elderly speech data and cognitive ability data and perform data perception to construct a sample of the elderly speech cognitive ability characteristics; The robust deep random configuration network construction module (2) is used to construct an integrated robust deep random configuration network model in a node increment manner, wherein the network model is composed of a plurality of base learners, and the learners include a plurality of hidden layers; The elderly cognitive ability prediction module (3) is used to perform integrated robust prediction on the speech cognitive ability feature samples according to the network model to obtain the network model integrated prediction result.

10. The system for predicting cognitive ability of the elderly based on speech tasks according to claim 9 is characterized in that: The robust deep random configuration network construction module (2) includes a weighted matrix calculation module (21) and a robust prediction module (22); The weighting matrix calculation module (21) is used to construct a predictor based on the hidden layer, adopt a kernel density estimation method and construct a residual probability density function based on the speech recognition ability feature sample, calculate and obtain a sample residual matrix of the predictor, and obtain a weighting matrix of the base learner according to the sample residual matrix; The robust prediction module (22) is used to adopt a weighted negative correlation learning method and perform integrated robust prediction based on the weighted matrix to obtain a network model output result.

Images