1. The utility model provides a mental health intelligence screening system based on BP neural network which characterized in that: the mental disease questionnaire training system comprises a mental disease questionnaire setting module, an investigation data evaluation and labeling module, a BP neural network model training module and a mental disease screening module which are sequentially connected; wherein the content of the first and second substances,

the mental disease questionnaire setting module is used for setting the contents of the mental disease questionnaire: the questionnaire takes different mental diseases as a unit, and there are 4 types of mental diseases in total, namely depression, generalized anxiety, mania and panic disorder; each disease detection corresponds to different problem numbers, wherein e problems exist in depression, f problems exist in generalized anxiety, g problems exist in mania, and h problems exist in panic disorder;

the survey data evaluation and labeling module evaluates and labels questionnaire survey results, and specifically comprises the following steps: after acquiring N pieces of questionnaire survey data, a professional psychologist evaluates each data sample one by one, and marks a label, namely, the type of a disease or no disease is marked;

the BP neural network model training module inputs data samples from the survey data evaluation and labeling module and performs model training to obtain a BP neural network model, and the BP neural network model training module specifically comprises the following steps: all N data samples were divided into 70% training set, 15% validation set and 15% testing set, training a single hidden layer BP neural network, wherein the input layer of the single hidden layer BP neural network is provided with e + f + g + h nodes, the input values of the three-dimensional data respectively correspond to the results of e problems of depression detection, f problems of generalized anxiety disorder detection, g problems of mania detection and h problems of panic disorder detection, the output layer has 16 nodes, the output value is the prediction result of the potential patients with the 4 types of diseases, the hidden layer has 17 neurons in total, a sigmoid function f (x) is 1/(1+ e ^ -x) (-x is a power number) is adopted as an activation function, the method comprises the steps that an activation function is not used in an output layer, a verification set is used for verifying a trained model, and finally a BP neural network model is determined after a test set test is passed;

the mental disease screening module carries out questionnaire survey on a person to be evaluated, data of the questionnaire survey are input into a BP neural network model obtained by a BP neural network model training module, and the illness condition of the person to be evaluated is judged, and the mental disease screening module specifically comprises the following steps:

filling a questionnaire by a person to be evaluated, and completing depression detection on e problems, general anxiety on f problems, mania on g problems and panic disorder on h problems in sequence to obtain e + f + g + h results;

and taking the obtained e + f + g + h results as the input of the BP neural network model to obtain the output value of the model, and judging whether the person to be evaluated suffers from one or more mental diseases of depression, generalized anxiety disorder, mania and panic disorder according to the output value.

Technical Field

At present, the peak of the 3.0 development of artificial intelligence is in progress, and the software and hardware capabilities of a computer and the artificial intelligence technology are enough to support the background of realizing intelligent medical treatment. The screening of psychosis for a wide range of populations at home and abroad mainly depends on the Mental Health screening scales such as Minnesota multipersonality testing Scale (MMPI), SCL-90 Symptom self-Rating Scale (Symptom Check List-90, SCL-90), mood disorder assessment Scale, Young's Manic Rating Scale (YMRS), Hamilton Depression Scale (Hamilton Depression Scale, HAMD), Mental Health diagnosis Test (MHT), and the like. Most of the screening scales are based on the symptoms of depressed mood, lack of interest or quick sense, insomnia or excessive sleep, sudden weight drop or rising, exhaustion and hypodynamia and the like in recent life. However, the screening scale-based mode is based on self-report of an evaluation object, is influenced by personal subjective belief and self-perception capability, subsequent clinical interviews depend on oral expressive ability of the evaluation object and subjective judgment of clinicians, and answers of the evaluation object are generally based on in the evaluation process, and are supplemented with multiple information such as facial expressions, limb behaviors and the like. Secondly, because the popularization and promotion strength of the mental health knowledge is not enough, the understanding rate of common residents in China on the mental health knowledge and the proportion of correctly selecting a medical mode are obviously low. Most people do not have basic understanding and understanding on mental diseases, lack understanding and concordance on patients, and have serious prejudice and discrimination phenomena, so that mental disease patients and families thereof generally have strong 'disease pubic feeling', or have misunderstanding and conflict on mental disease, misunderstanding and contradiction on to mental medical treatment, unwillingness or dare to receive diagnosis, treatment and psychological intervention, and delay treatment, so that the condition of illness is aggravated.

Disclosure of Invention

The invention aims to overcome the defects of the prior art and provides a mental health intelligent screening system based on a BP neural network, which reduces the artificial subjectivity of the traditional mental disease judgment method and completes the screening of new potential mental diseases of patients.

In consideration of mental health mind and actual use scenes of residents, the invention designs and develops the mental health intelligent screening system based on the BP neural network, so that the residents can carry out mental health detection privately and early discovery and early intervention are realized, and then the subjective judgment of clinicians is not depended on through the BP neural network, so that the mental disease screening is more objective and scientific.

The BP neural network essentially realizes a mapping function from input to output, is particularly suitable for solving the problem of complex internal mechanism, has high self-learning and self-adapting capability, and is very suitable for screening mental diseases.

In order to realize the technical task, the invention adopts the following technical solution:

the utility model provides a mental health intelligence screening system based on BP neural network which characterized in that: the mental disease examination and diagnosis system comprises a mental disease questionnaire setting module, an examination data evaluation and labeling module, a BP neural network model training module and a mental disease screening module which are sequentially connected; wherein the content of the first and second substances,

the mental disease questionnaire setting module is used for setting the contents of the mental disease questionnaire: questionnaires with different mental disorders as a unit, there are 4 categories of mental disorders in total, depression, generalized anxiety, mania, panic disorder, respectively; each disease detection corresponds to different problem numbers, wherein e problems exist in depression, f problems exist in generalized anxiety, g problems exist in mania, and h problems exist in panic disorder;

the survey data evaluation and labeling module evaluates and labels questionnaire survey results, and specifically comprises the following steps: after acquiring N pieces of questionnaire survey data, a professional psychologist evaluates each data sample one by one, and marks a label, namely, the type of a disease or no disease is marked;

the BP neural network model training module inputs data samples from the survey data evaluation and label module and performs model training to obtain a BP neural network model, and the BP neural network model training method specifically comprises the following steps: dividing all N data samples into a training set of 70%, a validation set of 15% and a test set of 15%, training a single-hidden-layer BP neural network, wherein the input layer of the single-hidden-layer BP neural network has e + f + g + h nodes, the input values of the nodes respectively correspond to the results of e problems of depression detection, f problems of generalized anxiety detection, g problems of mania detection and h problems of panic disorder detection, the output layer has 16 nodes, the output value of the node is the prediction result of potential patients suffering from the 4 types of diseases, the hidden layer has 17 neurons in total, a sigmoid f x) number (1/(1 + e-x) (-x) is taken as an activation function, the validation set is used for validating the trained model without using the activation function at the output layer, finally, determining a BP neural network model after the test set passes the test;

the mental disease screening module carries out questionnaire survey on a person to be evaluated, data of the questionnaire survey are input into a BP neural network model obtained by a BP neural network model training module, and the illness condition of the person to be evaluated is judged, and the mental disease screening module specifically comprises the following steps:

filling a questionnaire by a person to be evaluated (a new potential patient), and completing e problems of depression detection, f problems of generalized anxiety, g problems of mania and h problems of panic disorder in sequence to obtain e + f + g + h results;

and taking the obtained e + f + g + h results as the input of the BP neural network model to obtain the output value of the model, and judging whether the person to be evaluated suffers from one or more mental diseases of depression, generalized anxiety disorder, mania and panic disorder according to the output value.

The invention has the advantages that:

1) the BP neural network is used for obtaining the disease result of the potential patient, so that the accuracy is high, and the artificial subjectivity of the traditional mental disease judgment method is reduced;

2) the invention enables the potential patient to autonomously judge the mental disease, thereby greatly protecting the privacy of the potential patient.

Drawings

Fig. 1 is a schematic structural diagram of a BP neural network in an embodiment of the present invention.

Fig. 2 is a system configuration diagram of an embodiment of the present invention.

Detailed Description

The technical scheme of the invention is further explained by combining the attached drawings.

The utility model provides a mental health intelligence screening system based on BP neural network which characterized in that: the mental disease examination and diagnosis system comprises a mental disease questionnaire setting module, an examination data evaluation and labeling module, a BP neural network model training module and a mental disease screening module which are sequentially connected; wherein the content of the first and second substances,

the mental disease questionnaire setting module is used for setting the contents of the mental disease questionnaire: questionnaires with different mental disorders as a unit, there are 4 categories of mental disorders in total, depression, generalized anxiety, mania, panic disorder, respectively; each disease detection corresponds to different problem numbers, wherein e problems exist in depression, f problems exist in generalized anxiety, g problems exist in mania, and h problems exist in panic disorder;

the survey data evaluation and labeling module evaluates and labels questionnaire survey results, and specifically comprises the following steps: after acquiring N pieces of questionnaire survey data, a professional psychologist evaluates each data sample one by one, and marks a label, namely, the type of a disease or no disease is marked;

the BP neural network model training module inputs data samples from the survey data evaluation and label module and performs model training to obtain a BP neural network model, and the BP neural network model training method specifically comprises the following steps: dividing all N data samples into a training set of 70%, a validation set of 15% and a test set of 15%, training a single-hidden-layer BP neural network, wherein the input layer of the single-hidden-layer BP neural network has e + f + g + h nodes, the input values of the nodes respectively correspond to the results of e problems of depression detection, f problems of generalized anxiety detection, g problems of mania detection and h problems of panic disorder detection, the output layer has 16 nodes, the output value of the node is the prediction result of potential patients suffering from the 4 types of diseases, the hidden layer has 17 neurons in total, a sigmoid f x) number (1/(1 + e-x) (-x) is taken as an activation function, the validation set is used for validating the trained model without using the activation function at the output layer, finally, determining a BP neural network model after the test set passes the test;

the mental disease screening module carries out questionnaire survey on a person to be evaluated, data of the questionnaire survey are input into a BP neural network model obtained by a BP neural network model training module, and the illness condition of the person to be evaluated is judged, and the mental disease screening module specifically comprises the following steps:

filling a questionnaire by a person to be evaluated, and completing e problems of depression detection, f problems of extensive anxiety, g problems of mania and h problems of panic disorder in sequence to obtain e + f + g + h results;

and taking the obtained e + f + g + h results as the input of the BP neural network model to obtain the output value of the model, and judging whether the person to be evaluated suffers from one or more mental diseases of depression, generalized anxiety disorder, mania and panic disorder according to the output value.

Questionnaires with different mental disorders as a unit, there are 4 categories of mental disorders in total, depression, generalized anxiety, mania, and panic disorder, respectively; each disease detection corresponds to different problem numbers, wherein e problems exist in depression, f problems exist in generalized anxiety, g problems exist in mania, and h problems exist in panic disorder;

there are 16 total possibilities of mental illness, including a: single disease, b: binary mixed disease, c: ternary mixed disease, d: quaternary mixed diseases;

in this embodiment, e is 18, f is 11, g is 9, and h is 16, that is, there are 18 problems in depression detection, 11 problems in generalized anxiety, 9 problems in mania, and 16 problems in panic disorder;

step 2: after acquiring N pieces of questionnaire survey data, for example, N is 3000, each piece of data is evaluated one by a professional psychologist, and the data is labeled respectively;

and step 3: all N data samples were divided into 70% training set, 15% validation set and 15% testing set, training a single hidden layer BP neural network, wherein the network input layer has e + f + g + h nodes, the input values of the three-dimensional data respectively correspond to the results of e problems of depression detection, f problems of generalized anxiety disorder detection, g problems of mania detection and h problems of panic disorder detection, the output layer has 16 nodes, the output value is the prediction result of the potential patients with the 4 types of diseases, the hidden layer has 17 neurons in total, a sigmoid function f (x) is 1/(1+ e ^ -x) (-x is a power number) is adopted as an activation function, the method comprises the steps that an activation function is not used in an output layer, a verification set is used for verifying a trained model, and finally a BP neural network model is determined after a test set test is passed;

in this embodiment, the BP neural network input layer has 54 nodes, and its input values respectively correspond to results of 18 problems in depression detection, 11 problems in generalized anxiety disorder detection, 9 problems in mania detection, and 16 problems in panic disorder detection, and the output layer has 16 nodes, and its output value is a prediction result of a potential patient suffering from the 4 types of diseases, and the hidden layer has 17 neurons in total;

and 4, step 4: filling a questionnaire by a person to be evaluated (a new potential patient), and completing depression detection on e problems, generalized anxiety f problems, mania g problems and panic disorder h problems in sequence to obtain e + f + g + h results;

in this example, the subject completed 18 depression tests, 11 generalized anxiety tests, 9 manic disorders, and 16 panic disorders in sequence, giving 54 results;

and 5: and taking the obtained e + f + g + h results as the input of the BP neural network model to obtain the output value of the model, and judging whether the potential patient suffers from one or more mental diseases of depression, generalized anxiety disorder, mania and panic disorder according to the output value.

In this embodiment, 54 results will be obtained as inputs to the BP neural network model.

The embodiments described in this specification are merely illustrative of implementations of the inventive concept and the scope of the present invention should not be considered limited to the specific forms set forth in the embodiments but rather by the equivalents thereof as would occur to those skilled in the art upon consideration of the present inventive concept.