A prediction system for the risk of depression and anxiety based on non-invasive measurement indicators

Abstract

The present invention belongs to the technical field of mental illness prediction, and specifically relates to a prediction system for the risk of depression combined with anxiety based on non-invasive measurement indicators. The prediction system of the present invention includes: an input module for inputting characteristic indicators, the characteristic indicators include intestinal flora information, and the intestinal flora information includes whether the following bacteria are contained or the abundance of the following bacteria: Romboutsia bacteria, Alloprevotella bacteria, Tyzzerella bacteria, Barnesiella bacteria; a prediction module for calculating the characteristic indicators through a prediction model to obtain a prediction result of whether depression combined with anxiety occurs; an output module for outputting the prediction result. The present invention realizes a system for predicting the risk of depression combined with anxiety using a limited number of non-invasive measurement indicators. The present invention has the advantages of high result accuracy, a small number of indicators, and all are non-invasive measurements, and has a good application prospect.

Description

Technical Field

[0001] The present invention belongs to the technical field of mental illness prediction, and specifically relates to a prediction system for the risk of depression combined with anxiety based on non-invasive measurement indicators. Background Art

[0002] Depressive disorders and anxiety disorders are both common mental disorders. Early screening and intervention of depression and anxiety symptoms are conducive to reducing the medical burden. In recent years, studies have shown that there is a close relationship between the composition and function of the intestinal microbial community and depression and anxiety. This field is called the "gut-brain axis" or "gut-brain axis". The impact of intestinal microorganisms on mental illnesses such as depression and anxiety is specifically reflected in the following aspects: 1. Microorganisms affect neurotransmitters and metabolites: Intestinal microorganisms can produce a variety of neuroactive substances, including neurotransmitters (such as serotonin, dopamine), short-chain fatty acids (such as butyrate) and vitamin B group. These substances can enter the brain through the blood, affect neurotransmission, and regulate emotions and cognitive functions. 2. Immune system regulation: Intestinal flora participates in regulating the maturity and balance of the immune system. Abnormal immune response is related to psychological diseases such as depression and anxiety. Intestinal microorganisms indirectly affect the inflammatory state of the brain by regulating the immune system, which is related to depressive symptoms. 3. Barrier function: Intestinal flora maintains the integrity of the intestinal barrier and prevents the penetration of harmful substances. Impaired barrier function may lead to inflammation, inflammatory response, affect brain function, and is related to depression and anxiety. 4. Stress pathways: Gut microbes communicate with the brain through vasovagal pathways (such as the vagus nerve) and endocrine pathways (such as the adrenocorticotropic hormone system), affecting stress responses and mood regulation.

[0003] Due to the connection between intestinal microorganisms and depression and anxiety, it is possible to diagnose mental illnesses such as depression or anxiety by detecting the types of intestinal flora. For example, the Chinese invention patent application "CN202311133005.7 A method for establishing a depression auxiliary identification tool based on intestinal flora translocation" established a method for auxiliary diagnosis of depression through intestinal flora. However, the composition of the intestinal flora is complex, including a large number of different types of microorganisms, and the cost of detecting all the microbial species in the intestinal flora is too high to be practical. Therefore, it is an important topic in this field to study the indicators in the intestinal flora that can best reflect depression and anxiety, and to find a combination of bacteria that can efficiently and accurately predict depression and anxiety.

[0004] With the rapid development of computer technology, the medical field has gradually entered the era of big data and machine learning. Using machine learning technology to establish clinical prediction models can deeply explore the value of data and play an indispensable role in effectively supporting clinical decision-making. Clinical prediction models can be divided into diagnostic models and prognostic models. Predictive diagnostic models can estimate the probability of an individual currently suffering from a certain disease through mathematical formulas, while prognostic models refer to the prediction of the incidence of complications, recurrence, and death in a certain period of time in the future after being diagnosed with a certain disease. At present, commonly used machine learning algorithms include logistic regression, decision trees, support vector machines, deep neural networks, etc. Different model algorithms have different applicability for different prediction tasks and prediction indicators. Therefore, how to choose a suitable model algorithm for a specific task is also an important issue.

[0005] In short, there is currently a lack of relevant diagnostic models for using intestinal flora to predict depression and anxiety. How to select efficient and accurate prediction indicators and choose appropriate model algorithms is still a problem that needs to be solved urgently. Summary of the invention

[0006] In view of the problems of the prior art, the present invention provides a prediction system for the risk of depression combined with anxiety based on non-invasive measurement indicators.

[0007] A prediction system for the risk of depression and anxiety based on non-invasive measurement indicators, including:

[0008] An input module, used to input characteristic indicators, wherein the characteristic indicators include intestinal flora information, wherein the intestinal flora information includes whether the following bacterial species are present or the abundance of the following bacterial species: Romboutsia, Alloprevotella, Tyzzerella, and Barnesiella;

[0009] A prediction module is used to calculate the characteristic index through a prediction model to obtain a prediction result of whether depression and anxiety occur;

[0010] An output module is used to output the prediction result.

[0011] Preferably, the intestinal flora information includes the abundance of Romboutsia bacteria in the intestinal flora, whether Alloprevotella bacteria are present, whether Tyzzerella bacteria are present, and whether Barnesiella bacteria are present.

[0012] Preferably, the intestinal flora information is obtained by detecting stool samples using 16S rRNA high-throughput sequencing technology.

[0013] Preferably, the abundance of Romboutsia bacteria is assigned as follows: when the relative abundance of Romboutsia bacteria is greater than or equal to 0.024, the abundance of Romboutsia bacteria is high, and a value of 1 is assigned; when the relative abundance of Romboutsia bacteria is less than 0.024, the abundance of Romboutsia bacteria is low, and a value of 0 is assigned.

[0014] Preferably, the characteristic indicators also include the perceived economic status and whether there are sleep problems.

[0015] Preferably, the perceived economic status and whether there are sleep problems are determined through a questionnaire survey.

[0016] Preferably, the algorithm of the prediction model is selected from a decision tree algorithm.

[0017] Preferably, the structure of the prediction model is:

[0018] The abundance of Romboutsia, the presence of Alloprevotella, the presence of Tyzzerella, the presence of Barnesiella, perceived economic status and sleep problems were used as independent variables, the presence of Alloprevotella as the target variable, and each node went to different leaf nodes according to the subject's node variable. A total of 8 decision pathways were generated, of which the predicted probability of subjects experiencing depression and anxiety in 3 pathways was >50%. The population characteristics were:

[0019] ①The presence of Alloprevotella bacteria;

[0020] ② There is no Alloprevotella, no Tyzzerella, but there is Barnesiella, and the abundance of Romboutsia is high;

[0021] ③ There is no Alloprevotella, no Tyzzerella, there is Barnesiella, there are sleep problems, the abundance of Romboutsia is high, and the perceived economic situation is average or poor.

[0022] Preferably, it also includes:

[0023] Database, used to store the data set used to train the model;

[0024] The model training module is used to call the data set and train the prediction model.

[0025] The present invention also provides a computer-readable storage medium on which a computer program for implementing the above prediction system is stored.

[0026] In the present invention, the "combined depression and anxiety" refers to suffering from both depressive disorder and anxiety disorder.

[0027] On the one hand, the present invention studies intestinal flora data and other data (such as sociodemographic information data, dietary intake data, lifestyle data), and selects the index combination with the best predictive performance for depression and anxiety. In the best embodiment, this combination is "conscious economic status, sleep problems, Romboutsia flora, Alloprevotella flora, Tyzzerella flora, Barnesiella flora". This index combination has the advantage of being able to accurately predict depression and anxiety with a small number of indicators. And the above indicators are all indicators that can be obtained by non-invasive detection. On the other hand, the present invention selects the best model algorithm based on the above optimized indicators.

[0028] In summary, the present invention does not require a large number of examinations and scales, and can accurately predict the risk of depression and anxiety disorders in subjects through limited non-invasive testing, which is conducive to the early screening of depression and anxiety disorders, and also provides time for interventional treatment of depression and anxiety disorders, and has good application prospects.

[0029] Obviously, according to the above contents of the present invention, in accordance with common technical knowledge and customary means in the art, without departing from the above basic technical ideas of the present invention, other various forms of modification, replacement or change may be made.

[0030] The above contents of the present invention are further described in detail below through specific implementation methods in the form of embodiments. However, this should not be understood as the scope of the above subject matter of the present invention being limited to the following examples. All technologies realized based on the above contents of the present invention belong to the scope of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] Figure 1 A schematic diagram of the prediction model constructed in Example 1;

[0032] Figure 2 The ROC curve of the prediction model constructed in Example 1;

[0033] Figure 3 It is a schematic diagram of the comparative prediction model 1 in Experimental Example 1;

[0034] Figure 4 It is the ROC curve of the comparison prediction model 1 in experimental example 1;

[0035] Figure 5 It is a schematic diagram of the comparative prediction model 2 in Experimental Example 1;

[0036] Figure 6This is the ROC curve of the comparison prediction model 2 in experimental example 1. DETAILED DESCRIPTION

[0037] It should be noted that the algorithms of data collection, transmission, storage and processing steps not specifically described in the embodiments, as well as the hardware structure, circuit connection, etc. not specifically described can all be implemented through the contents disclosed in the prior art.

[0038] Example 1 Prediction system for the risk of depression and anxiety based on non-invasive measurement indicators

[0039] The prediction system of this embodiment includes:

[0040] An input module, used to input characteristic indicators, wherein the characteristic indicators include intestinal flora information and non-intestinal flora information, wherein the intestinal flora information includes the abundance of Romboutsia bacteria in the intestinal flora, whether Alloprevotella bacteria are present, whether Tyzzerella bacteria are present, and whether Barnesiella bacteria are present; and the non-intestinal flora information includes the perceived economic status and whether there are sleep problems;

[0041] A prediction module is used to calculate the characteristic index through a prediction model to obtain a prediction result of whether depression and anxiety occur;

[0042] An output module, used for outputting the prediction result;

[0043] Database, used to store the data set used to train the model;

[0044] The model training module is used to call the data set and train the prediction model.

[0045] In this embodiment, the algorithm of the prediction model is preferably the ID3 decision tree algorithm, and the assignment method is:

[0046]

[0047]

[0048] The abundance of Romboutsia bacteria is assigned as follows: when the relative abundance of Romboutsia bacteria is greater than or equal to 0.024, the abundance of Romboutsia bacteria is high, and a value of 1 is assigned; when the relative abundance of Romboutsia bacteria is less than 0.024, the abundance of Romboutsia bacteria is low, and a value of 0 is assigned.

[0049] The structure of the prediction model is as follows Figure 1 As shown, specifically:

[0050] The abundance of Romboutsia, the presence of Alloprevotella, the presence of Tyzzerella, the presence of Barnesiella, perceived economic status and sleep problems were used as independent variables, the presence of Alloprevotella as the target variable, and each node went to different leaf nodes according to the subject's node variable. A total of 8 decision pathways were generated, of which the predicted probability of subjects experiencing depression and anxiety in 3 pathways was >50%. The population characteristics were:

[0051] ①The presence of Alloprevotella bacteria;

[0052] ② There is no Alloprevotella, no Tyzzerella, but there is Barnesiella, and the abundance of Romboutsia is high;

[0053] ③ There is no Alloprevotella, no Tyzzerella, there is Barnesiella, there are sleep problems, the abundance of Romboutsia is high, and the perceived economic situation is average or poor.

[0054] The technical solution of the present invention is further illustrated by experiments below.

[0055] Experimental Example 1 Modeling process and prediction performance comparison of different indicator combinations

[0056] 1.1 Methods

[0057] Subjects: In September 2023, cluster sampling was used to select adult males from Sichuan Province as research subjects. Inclusion criteria: (1) aged 18 to 65 years; (2) with a certain level of education, able to read the questionnaire independently and complete the questionnaire under guidance; (3) voluntarily signed the informed consent form; Exclusion criteria: (1) those with a history of drug abuse or drug crime; (2) those with a history of infectious diseases (such as AIDS, hepatitis C, etc.); (3) those with symptoms of alcohol or drug addiction; (4) those with a history of psychotropic drug use in the past month (except for the use of sleeping pills); (5) those with a history of brain trauma and coma for more than (1) Those who have had a history of urinary incontinence or vomiting for 10 minutes; (2) Those with neurological diseases (previously diagnosed with cerebral hemorrhage, cerebral infarction, Alzheimer's disease, Parkinson's disease, etc.); (3) Those with severe mental disorders (previously diagnosed with schizophrenia, bipolar disorder, mental retardation, dementia, etc.); (4) Those with a history of cancer; (5) Those who have undergone major gastrointestinal surgery within 5 years (such as cholecystectomy, appendectomy, gastrectomy, partial intestinal resection, etc.); (6) Those who have used antibiotics or probiotic preparations in the past month.

[0058] Socio-demographic information: Socio-demographic information was obtained through questionnaire survey and filled in by the subjects themselves, including name, date of birth (accurate to month), marital status (married / divorced / widowed / single), education level (primary school and below / junior high school / senior high school or technical secondary school / college and above), residential area (rural / urban), occupation (full-time / part-time / unemployed), perceived economic situation (good / better / average / poor), monthly family income (below 6000 / 6000-12000 / above 12000), etc.

[0059] Depression and anxiety assessment: Depression status was collected using the PHQ-9 scale, and anxiety status was collected using the GAD-7 scale. Both scales have good reliability and validity in assessing depression or anxiety.

[0060] Dietary survey: A Food Frequency Questionnaire (FFQ) that has been validated and verified among adults in the southwest region was used to collect data on the frequency and average intake of different food categories in the past year from 200 men, including refined grains, coarse grains, vegetables, fruits, livestock and poultry meat, aquatic products, eggs, pickled foods, beans and bean products, milk and dairy products, beverages, etc. The frequency of food intake was set to four units: times / day, times / week, times / month, and times / year.

[0061] Lifestyle status: A questionnaire was used to collect lifestyle status, including smoking history, drinking history, and sleep problems.

[0062] Distribution of intestinal flora: Fecal sample collection and DNA extraction: Qualified researchers were provided with fecal collection kits and instructed to read the collection instructions on the outer packaging of the collection device. The subjects collected samples according to the collection instructions by natural defecation. The genomic DNA of the samples was extracted by the sodium dodecyl sulfate (SDS) method, and the purity and concentration of the DNA were detected by agarose gel electrophoresis. The fecal 16S rRNA gene V4 region was amplified by PCR using primers 515F (SEQ ID NO.1: 5′-GTGCCAGCMGCCGCGGTAA-3′) and 806R (SEQ ID NO.2: 5′-GGACTACHVGGGTWTCTAAT-3′), and high-throughput sequencing was completed.

[0063] Construction of prediction model: Based on the set.seed function and the sample function, 200 subjects were divided into training set (Training Data) and test set (Test Data) in a ratio of 7:3 by random number sampling. The training set was used for further selection of prediction indicators and decision tree model fitting, while the test set was used for model evaluation. The rpart package and rpart.plot package in R 4.3.2 software were used to realize the construction and graphics generation of decision trees. The ID3 decision tree algorithm was used to construct a predictive diagnostic model for the occurrence of depression and anxiety in adult men from the aspects of sociodemographics, dietary intake, lifestyle and intestinal flora. The decision tree graphics were drawn using Microsoft Visio. The predict function was used to predict the classification of the observed indicators in the test set. The sensitivity and specificity were calculated by the confusion matrix. The AUC value was obtained using the ROC curve. The prediction effect of the established decision tree was analyzed. The plotROC function of the InformationValue R package was used to draw the ROC curve.

[0064] 1.2 Results

[0065] Subject data information:

[0066] A total of 200 male subjects completed all relevant examinations, of which 54 were diagnosed with anxiety or depression, accounting for 27.0%;

[0067] All subjects were randomly assigned in a 7:3 sampling ratio, with 140 and 60 subjects allocated to the training set and validation set, respectively.

[0068] There are 38 relevant feature indicators that can be used for modeling, as shown in Table 1 and Table 2:

[0069] Table 1: Characteristic indicators of non-intestinal flora

[0070]

[0071]

[0072] Table 2: Intestinal flora related signs and symptoms

[0073]

[0074] Among them, the judgment thresholds for low relative abundance and high relative abundance were the median relative abundance of the bacterial community, among which Firmicutes was 0.735, Bacteroidota was 0.191, Proteobacteria was 0.009, Bacteroides was 0.084, Romboutsia was 0.024, and Coprococcus was 0.005.

[0075] By optimizing the above-mentioned characteristic indicators, some characteristic indicators are selected for modeling the prediction model.

[0076] 1.2.1 Prediction Model of Example 1

[0077] According to the six characteristic indicators selected in Example 1 (self-perceived economic status, sleep problems, Romboutsia, Alloprevotella, Tyzzerella, and Barnesiella) as independent variables, whether depression and anxiety occur as the dependent variable, a predictive diagnosis model for whether adult males are depressed is constructed based on the ID3 decision tree algorithm. The modeling results are shown in Figure 1 shown.

[0078] Validation of the prediction model:

[0079] At present, the main evaluation indicators of the prediction model's discriminative ability include the area under the receiver operating characteristic curve (AUC), sensitivity, and specificity. Sensitivity refers to the ratio of positive cases to the total number of positive cases that the prediction model can correctly predict, reflecting the model's ability to screen positive patients. Specificity refers to the ratio of negative cases to the total number of negative cases that the prediction model can correctly predict, reflecting the model's ability to identify negative patients and exclude them. The AUC value is used as a comprehensive indicator to evaluate the accuracy of the prediction model, with a value range of 0.5 to 1. The larger the value, the higher the model's discriminative ability.

[0080] The predict function in R 4.3.2 software was used to predict the classification of the observed indicators in the test set. The sensitivity and specificity were calculated by the confusion matrix. The ROC curve was used to obtain the AUC value. The prediction effect of the established decision tree was analyzed. The plotROC function of the InformationValue R package was used to draw the ROC curve. For the established prediction model, the test set (accounting for 30% of the total data) was used to measure the performance of the established decision tree model. The results are shown in Figure 2. Figure 2 The sensitivity of the model was 0.852, the specificity was 0.630, and the AUC value was 0.803. The results showed that the depression and anxiety disorder prediction model constructed in Example 1 had good prediction ability.

[0081] 1.2.2 Comparison of prediction models 1

[0082] A predictive diagnostic model for depression in adult males was constructed based on the ID3 decision tree algorithm. Age, smoking history, sleep problems, livestock and poultry meat intake, milk and dairy product intake, Tyzzerella bacteria, and Paraprevotella bacteria were used as independent variables, and whether depression and anxiety occurred was used as the dependent variable set for modeling.

[0083] The prediction model structure is as follows Figure 3 The prediction performance results are shown in Figure 4 As shown in the figure, the sensitivity of the model is 0.791, the specificity is 0.629, and the AUC value is 0.753. Compared with Example 1, the prediction ability of the comparative prediction model is poor.

[0084] 1.2.3 Comparison of prediction models 2

[0085] A predictive diagnostic model for depression in adult males was constructed based on the ID3 decision tree algorithm, with occupation, perceived economic status, sleep problems, Alloprevotella, Tyzzerella, and Barnesiella as independent variables and depression combined with anxiety as the dependent variable set for modeling.

[0086] The prediction model structure is as follows Figure 5 The prediction performance results are shown in Figure 6 As shown in the figure, the sensitivity of the model is 0.642, the specificity is 0.824, and the AUC value is 0.776. Compared with Example 1, the prediction ability of the comparative prediction model is poor.

[0087] 1.2.4 Comparison of prediction models 3-22

[0088] In order to further prove that the indicators selected in the prediction model of Example 1 are the best combination, the dominant bacteria (Coprococcus bacteria, Firmicutes bacteria, Bacteroidota bacteria, Proteobacteria bacteria, Bacteroides bacteria) of the experimental population are used to replace the bacteria in the indicators of the prediction model of Example 1, and the prediction performance of these comparative prediction models is compared.

[0089] Specifically, any one of the five indicators, namely, the abundance of Coprococcus in the intestinal flora, the abundance of Firmicutes in the intestinal flora, the abundance of Bacteroidota in the intestinal flora, the abundance of Proteobacteria in the intestinal flora, and the abundance of Bacteroides in the intestinal flora, is used to replace any one of the flora-related indicators in the prediction model of Example 1. The replacement method is shown in the following table:

[0090] Table 3 Overview of comparison models

[0091]

[0092] The details of each model are as follows:

[0093] Table 4 Best model of Example 1

[0094]

[0095]

[0096] Table 5 Comparison of prediction model 3 (Coprococcus bacteria replace Romboutsia bacteria)

[0097]

[0098] Table 6 Comparison of prediction model 4 (Coprococcus bacteria replace Barnesiella bacteria)

[0099]

[0100] Table 7 Comparison of prediction model 5 (Coprococcus bacteria replaces Tyzzerella bacteria)

[0101]

[0102]

[0103] Table 8 Comparison of prediction model 6 (Coprococcus bacteria replaces Alloprevotella bacteria)

[0104]

[0105] Table 9 Comparison of prediction model 7 (Firmicutes bacteria replace Romboutsia bacteria)

[0106]

[0107] Table 10 Comparison of prediction model 8 (Firmicutes bacteria replace Barnesiella bacteria)

[0108]

[0109] Table 11 Comparison of prediction model 9 (Firmicutes bacteria replace Tyzzerella bacteria)

[0110]

[0111] Table 12 Comparison of prediction model 10 (Firmicutes bacteria replace Alloprevotella bacteria)

[0112]

[0113] Table 13 Comparison of prediction model 11 (Bacteroidota bacteria replace Romboutsia bacteria)

[0114]

[0115] Table 14 Comparison of prediction model 12 (Bacteroidota bacteria replace Barnesiella bacteria)

[0116]

[0117]

[0118] Table 15 Comparison of prediction model 13 (Bacteroidota bacteria replace Tyzzerella bacteria)

[0119]

[0120] Table 16 Comparison of prediction model 14 (Bacteroidota bacteria replace Alloprevotella bacteria)

[0121]

[0122] Table 17 Comparison of prediction model 15 (Proteobacteria bacteria replace Romboutsia bacteria)

[0123]

[0124]

[0125] Table 18 Comparison prediction model 16 (Proteobacteria bacteria replace Barnesiella bacteria)

[0126]

[0127] Table 19 Comparison of prediction model 17 (Proteobacteria bacteria replace Tyzzerella bacteria)

[0128]

[0129] Table 20 Comparison of prediction model 18 (Proteobacteria bacteria replace Alloprevotella bacteria)

[0130]

[0131]

[0132] Table 21 Comparison of prediction model 19 (Bacteroides bacteria replace Romboutsia bacteria)

[0133]

[0134] Table 22 Comparison of prediction model 20 (Bacteroides bacteria replace Barnesiella bacteria)

[0135]

[0136] Table 23 Comparison of prediction model 21 (Bacteroides bacteria replace Tyzzerella bacteria)

[0137]

[0138] Table 24 Comparison of prediction model 22 (Bacteroides bacteria replace Alloprevotella bacteria)

[0139]

[0140] Pseudo R2 is an indicator for evaluating the degree of fit of the regression model. The closer it is to 1, the better the model performance. By comparing the performance of the above-mentioned comparative prediction model with the prediction model of Example 1, it was found that replacing any one of the bacteria in the prediction model of Example 1 with other dominant bacteria in the experimental population would lead to a decrease in the model fit. This shows that the intestinal flora-related indicators selected by the present invention are the best indicator combination for predicting the risk of depression and anxiety.

[0141] Through the above embodiments and experimental examples, it can be seen that the present invention optimizes the combination of characteristic indicators and the modeling algorithm, and realizes a system for predicting the risk of depression and anxiety using a limited number of non-invasive measurement indicators. The present invention has the advantages of high result accuracy, a small number of indicators, and all non-invasive measurements, and has a good application prospect.

Claims

1. A prediction system for the risk of depression and anxiety based on non-invasive measurement indicators, characterized in that: include: An input module is used to input characteristic indicators, wherein the characteristic indicators include intestinal flora information, self-perceived economic status, and whether there are sleep problems. The intestinal flora information includes Romboutsia The abundance of bacteria and whether Alloprevotella Bacteria, whether it contains Tyzzerella Bacteria, whether it contains Barnesiella Bacteria; The self-perceived economic status and whether there are sleep problems are determined by questionnaire survey; A prediction module is used to calculate the characteristic index through a prediction model to obtain a prediction result of whether depression and anxiety occur; An output module, used for outputting the prediction result; The algorithm of the prediction model is selected from a decision tree algorithm; The structure of the prediction model is: by Romboutsia The abundance of bacteria and whether Alloprevotella Bacteria, whether it contains Tyzzerella Bacteria, whether it contains Barnesiella bacteria, perceived economic status and sleep problems were used as independent variables, and whether the subjects had depression and anxiety was used as the target variable. Alloprevotella Bacteria is the root node variable. Each node goes to different leaf nodes according to the subject node variable, thus generating 8 decision paths in total. Among them, the predicted probability of subjects suffering from depression and anxiety in 3 paths is >50%. The group characteristics are: ①Existence Alloprevotella bacteria; ② Does not exist Alloprevotella Bacteria, not present Tyzzerella Bacteria, exist Barnesiella bacteria, Romboutsia The abundance of bacteria is high; ③Does not exist Alloprevotella Bacteria, not present Tyzzerella Bacteria, exist Barnesiella bacteria, have sleeping problems, Romboutsia The abundance of bacteria is high, and the perceived economic status is average or poor; Said Romboutsia The abundance of bacteria is assigned as follows: Romboutsia The relative abundance of bacteria is greater than or equal to 0.

024. Romboutsia The abundance of bacteria is high, and the value is 1; Romboutsia The relative abundance of bacteria is less than 0.

024. Romboutsia The abundance of bacteria is low and is assigned a value of 0.

2. The prediction system according to claim 1, characterized in that: The intestinal flora information is obtained by detecting fecal samples using 16S rRNA high-throughput sequencing technology.

3. The prediction system according to claim 1, characterized in that Also includes: Database, used to store the data set used to train the model; The model training module is used to call the data set and train the prediction model.

4. A computer-readable storage medium, characterized in that: A computer program for implementing the prediction system according to any one of claims 1 to 3 is stored thereon.

Images