Malignant tumor risk assessment and diagnosis decision system, equipment and medium

Abstract

The invention relates to a malignant tumor risk assessment and diagnosis decision-making system and device and a medium, and relates to the technical field of malignant tumor assessment, and the malignant tumor risk assessment and diagnosis decision-making system comprises an information acquisition module used for acquiring user information, a data processing module used for processing user data in the user information to obtain user verification data, and a feature extraction module used for extracting features from the user verification data. The feature extraction module is used for carrying out data feature extraction on user verification data to obtain data parameter features corresponding to a user and parameter feature values corresponding to the data parameter features, and the feature evaluation module is used for inputting the data parameter features and the parameter feature values into a prediction model for training. And the diagnosis feedback module is used for carrying out format conversion on the disease probability values based on a preset conversion standard to obtain disease probability information and feeding back the disease probability information to the target equipment. The malignant tumor risk assessment accuracy is improved.

Description

Technical Field

[0001] The present application relates to the technical field of tumor assessment, and in particular to a malignant tumor risk assessment and diagnosis decision system, equipment and medium. Background Art

[0002] Malignant tumors are becoming more and more important as a factor affecting the physical health of residents. Due to their rapid development and harmfulness, malignant tumors pose a serious threat to the physical health of residents. Effective malignant tumor risk assessment and screening can ensure timely detection of cancer and early treatment.

[0003] At present, the malignant tumor risk assessment method is often used to assess the risk of malignant tumors of resident users, that is, by collecting the physical information parameters of resident users and inputting the physical information parameters into the risk assessment model for training, the user's assessment score is determined using the trained risk assessment model. When the assessment score exceeds the preset score, it means that the current user has a risk of a certain malignant tumor. If the assessment score does not exceed the preset score, it means that the current user does not have a risk of malignant tumor. When training the risk assessment model, a large amount of historical malignant tumor disease data is often used for training, and the weights of the disease characteristics of different malignant tumors are calculated to determine the assessment scores corresponding to different disease characteristics.

[0004] However, through analyzing historical patient data, this application has learned that the weights corresponding to the disease characteristics of different malignant tumors are different in different time periods. This means that people in different eras have different living habits, and the proportion of factors that ultimately lead to the onset of malignant tumors is also different. Therefore, if the risk of a user suffering from a malignant tumor is judged only based on the evaluation score obtained by the risk assessment model, there is a defect of reducing the accuracy of risk assessment. Summary of the invention

[0005] In order to improve the accuracy of malignant tumor risk assessment, the present application provides a malignant tumor risk assessment and diagnosis decision system, equipment and medium.

[0006] In the first aspect, the present application provides a malignant tumor risk assessment and diagnosis decision system, which adopts the following technical solutions: A malignant tumor risk assessment and diagnosis decision system, comprising: An information acquisition module, used to acquire user information, wherein the user information is personal information input by the user through the target device; A data processing module, used to process the user data in the user information to obtain user verification data; A feature extraction module, used to extract data features from the user verification data to obtain data parameter features corresponding to the user and parameter feature values ​​corresponding to the data parameter features; A feature evaluation module, used for inputting the data parameter features and the parameter feature values ​​into a prediction model for training, and obtaining the disease probability values ​​of the user at different time stages; The diagnosis feedback module is used to convert the format of the disease probability value based on a preset conversion standard to obtain disease probability information, and feed back the disease probability information to the target device.

[0007] In another possible implementation, the system further includes: a feature data acquisition module, a data determination module, a data sorting module, a data analysis module and a model training module, wherein: The characteristic data acquisition module is used to acquire the characteristics of malignant tumor factors corresponding to different types of malignant tumors and the historical data of patients suffering from different types of malignant tumors; The data determination module is used to determine the factor characteristic data corresponding to each malignant tumor factor characteristic and the time series length corresponding to the factor characteristic data according to the patient historical data; The data sorting module is used to perform unsupervised time series data sorting on the patient historical data according to the factor feature data and the time series length to obtain patient data matrices corresponding to different types of malignant tumors; The data analysis module is used to perform data analysis on the patient data matrix to obtain the weight value distribution of different malignant tumor factor characteristics within a preset time in the future and the weight scores corresponding to the weight value distribution; The model training module is used to create a prediction model, and input the malignant tumor factor characteristics, weight value distribution and weight score into the prediction model for training to obtain a trained prediction model.

[0008] In another possible implementation, when the data analysis module performs data analysis on the patient data matrix to obtain the weight value distribution of different malignant tumor factor characteristics within a preset future time and the weight score corresponding to the weight value distribution, it is specifically used to: Determine whether the malignant tumor factor feature is an uncontrollable factor feature, and if so, extract and define the data corresponding to the malignant tumor factor feature in the patient data matrix as the first sequence data, and define the remaining data corresponding to the malignant tumor factor feature in the patient data matrix as the second sequence data; Calculate the data development corresponding to each malignant tumor factor feature in the first sequence data and the second sequence data according to the time series monitoring, and obtain the free association coefficient corresponding to the first sequence data and the controllable association coefficient corresponding to the second sequence data; Calculating the weight value of each malignant tumor factor feature in the first sequence data in different unit times based on the free association coefficient and the preset weight standard to obtain a first time weight group of different malignant tumor factor features in the first sequence data; Calculating the weight value of each malignant tumor factor feature in the second sequence data in different unit times based on the controllable correlation coefficient and the preset weight standard to obtain a second time weight group of different malignant tumor factor features in the second sequence data; Respectively performing data correspondence and combining processing on the first time weight group and the second time weight group with the patient data matrix to generate a prediction data matrix; The prediction data matrix is ​​processed to obtain the weight value distribution of different malignant tumor factor characteristics within a preset future time and the weight scores corresponding to the weight value distribution.

[0009] In another possible implementation, when the data analysis module processes the prediction data matrix to obtain the weight value distribution of different malignant tumor factor characteristics within a preset future time and the weight score corresponding to the weight value distribution, it is specifically used to: Performing data processing on the prediction data matrix to obtain a first training matrix; The first training matrix is ​​input into a preset algorithm model for data extrapolation to obtain the weight value distribution of different malignant tumor factor characteristics within a future preset time and the weight score corresponding to the weight value distribution.

[0010] In another possible implementation, when the data analysis module performs data correspondence and combination processing on the first time weight group and the second time weight group and the patient data matrix to generate a prediction data matrix, the data analysis module is specifically used to: respectively integrating the first time weight group and the second time weight group with the patient data matrix to generate an integrated data matrix; Performing basic data distribution exploration on the integrated data matrix to obtain the relative periodicity of the weight value changes of the different malignant tumor factor characteristics within a unit time, and determining the time period length based on the relative periodicity; Performing supervised time series data sorting on the integrated data matrix based on the time period length to obtain a sorted data matrix; Based on the sorted data matrix, the change trend of the weight value group within a future preset time period is predicted to generate prediction matrix data.

[0011] In another possible implementation manner, when the data analysis module performs data processing on the prediction data matrix to obtain the first training matrix, it is specifically used to: Calculating a normal distribution mean and a normal distribution variance of the feature data contained in the prediction data matrix, and determining a 3σ range of the prediction data matrix based on the normal distribution mean and the normal distribution variance; Determine whether the feature data is outside the 3σ range, if the feature data is outside the 3σ range, determine a first matrix sequence of the prediction data matrix where the feature data is located, calculate a sequence average value according to the first matrix sequence, replace the feature data with the sequence average value, obtain a replaced second matrix sequence, and perform missing value processing on the second matrix sequence; A sequence normalization process is performed on the second matrix sequence in the prediction data matrix to obtain a first training matrix.

[0012] In another possible implementation, the system further includes: a weight determination module and a weight reset module, wherein: The weight determination module is used to determine the actual weight distribution and actual weight score of the different malignant tumor factor characteristics detected within a future preset time period; The weight resetting module is used to determine whether the actual weight distribution and the actual weight score are consistent with the weight value distribution and the weight score respectively. If they are inconsistent, the weight distribution and the weight score are denormalized based on the actual weight distribution and the actual weight score to restore the weight distribution and the weight score to the actual weight distribution and the actual weight score.

[0013] In another possible implementation, the system further includes: a disease determination module and a suggestion generation module, wherein: The disease determination module is used to determine the disease characteristics of the user and the risk level corresponding to the disease characteristics based on the disease probability information; The suggestion generation module is used to input the disease characteristics and risk levels into the trained prevention model to obtain prevention suggestion information, and send the prevention suggestion information to the target device.

[0014] In a second aspect, the present application provides an electronic device, which adopts the following technical solution: An electronic device, comprising: one or more processors; Memory; One or more applications, wherein one or more applications are stored in a memory and configured to be executed by one or more processors, and the one or more programs are configured to: execute a malignant tumor risk assessment and diagnosis decision system shown in any possible implementation of the first aspect.

[0015] In a third aspect, the present application provides a computer-readable storage medium, which adopts the following technical solution: A computer-readable storage medium includes: a computer program that can be loaded and executed by a processor to implement a malignant tumor risk assessment and diagnosis decision system shown in any possible implementation of the first aspect.

[0016] In summary, the present application includes at least one of the following beneficial technical effects: By adopting the above technical solution, when conducting a malignant tumor susceptibility assessment for a user, the user personal information input by the user into the target device is obtained, and then the user data in the user information is processed to obtain user verification data, and data features are extracted from the user verification data to obtain data parameter features corresponding to the user and parameter feature values ​​corresponding to the data parameter features. The data parameter features and parameter feature values ​​are then input into a prediction model for training to obtain the disease probability values ​​of the user at different time stages, and then the disease probability values ​​are formatted based on a preset conversion standard to obtain disease probability information, and the disease probability information is fed back to the target device. By viewing the information fed back by the target device, the user can know the disease probability values ​​corresponding to the user at different time stages, thereby achieving the effect of improving the accuracy of malignant tumor risk assessment. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 It is a schematic diagram of the structure of a malignant tumor risk assessment and diagnosis decision system according to an embodiment of the present application; Figure 2 It is a flowchart of a malignant tumor risk assessment and diagnosis decision system structure in an embodiment of the present application; Figure 3 It is a schematic diagram of the structure of an electronic device according to an embodiment of the present application. DETAILED DESCRIPTION

[0018] The following is combined with Figure 1-3 This application is described in further detail.

[0019] After reading this specification, those skilled in the art may make non-creative modifications to this embodiment as needed, but such modifications are protected by patent law as long as they are within the scope of the claims of this application.

[0020] In order to make the purpose, technical solution and advantages of the embodiments of the present application clearer, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0021] In addition, the term "and / or" in this article is only a description of the association relationship of associated objects, indicating that there can be three relationships. For example, A and / or B can represent: A exists alone, A and B exist at the same time, and B exists alone. In addition, the character " / " in this article, unless otherwise specified, generally means that the associated objects before and after are in an "or" relationship.

[0022] The embodiments of the present application are further described in detail below in conjunction with the drawings in the specification.

[0023] The embodiment of the present application provides a malignant tumor risk assessment and diagnosis decision system, which is executed by an electronic device, which can be a server or a terminal device, wherein the server can be an independent physical server, or a server cluster or distributed system composed of multiple physical servers, or a cloud server that provides cloud computing services. The terminal device can be a smart phone, a tablet computer, a laptop computer, a desktop computer, etc., but is not limited to this. The terminal device and the server can be directly or indirectly connected via wired or wireless communication, and the embodiment of the present application does not limit this. Figure 1-Figure 2 As shown, the system includes: an information acquisition module, a data processing module, a feature extraction module, a feature evaluation module and a diagnosis feedback module, wherein: The information acquisition module is used to obtain user information.

[0024] The user information is the personal information input by the user through the target device.

[0025] Specifically, the user inputs personal information through the target device, where the personal information includes age, gender, family medical history, lifestyle, etc. The target device includes the user's personal mobile / fixed terminal device and the fixed terminal device located in the hospital.

[0026] The data processing module is used to process the user data in the user information to obtain user verification data.

[0027] In the embodiment of the present application, in order to ensure the data integrity and accuracy of the user information input by the user, the user information is verified by data verification / cleaning to ensure that the data is correct. Specifically, data verification includes integrity verification, accuracy verification and consistency verification, among which: Completeness verification: Check whether the data is complete and whether necessary fields or attributes are missing. Accuracy verification: Verify whether the data conforms to the expected format, range or logical relationship. For example, check whether the date field conforms to the date format and check whether the numerical field is within a reasonable range. Consistency verification: Verify whether the data is consistent with the data in other data sources or databases. For example, check whether the values ​​of the same field in different data tables are the same. Data cleaning includes missing value processing, outlier processing, duplicate value processing, format conversion, and standardization and normalization. Among them, missing value processing: For missing data, you can consider filling default values, using the mean or median, or deleting rows or columns with missing values. Outlier processing: Outliers may have a negative impact on data analysis. Outliers can be detected and processed through statistical methods, visualization, or using machine learning algorithms. Duplicate value processing: Delete duplicate rows or columns, or merge duplicate data rows. Format conversion: Convert data from one format to another, such as converting text data to numerical data, or converting dates from one format to another. Standardization and normalization: Convert data to a standardized scale to eliminate scale differences between data.

[0028] The feature extraction module is used to extract data features from user verification data to obtain data parameter features corresponding to the user and parameter feature values ​​corresponding to the data parameter features.

[0029] Specifically, data features are extracted from the user verification data to obtain the user's data parameter features, where the data parameter features include: age characteristics, smoking history characteristics, drinking characteristics, occupational characteristics, family history characteristics, chest radiotherapy characteristics, BMI characteristics, and thyroid characteristics, etc., and the parameter feature values ​​are the data values ​​that correspond one-to-one to the data parameter features.

[0030] For the embodiments of the present application, when extracting data features, statistical methods, dimension reduction methods, and machine learning algorithms can be used, among which, statistical methods: extract features by counting the frequency or distribution of certain features in the data set. Dimension reduction method: extract the most useful features in the data by reducing the high-dimensional data to a low dimension. For example, dimension reduction methods such as PCA or LDA can be used. Machine learning algorithm: machine learning algorithms such as decision trees, neural networks, support vector machines, etc. can be used for feature selection and model training. By training the model, more in-depth features can be discovered.

[0031] The feature evaluation module is used to input data parameter features and parameter feature values ​​into the prediction model for training to obtain the user's disease probability values ​​at different time stages.

[0032] In order to obtain the prediction model, the system also includes: a feature data acquisition module, a data determination module, a data sorting module, a data analysis module and a model training module, wherein: The feature data acquisition module is used to obtain the malignant tumor factor features corresponding to different types of malignant tumors and the historical data of patients suffering from different types of malignant tumors.

[0033] The data determination module is used to determine the factor characteristic data corresponding to each malignant tumor factor characteristic and the time series length corresponding to the factor characteristic data according to the patient's historical data.

[0034] The data sorting module is used to sort the patient's historical data in an unsupervised time series according to the factor characteristic data and the time series length, and obtain the patient data matrix corresponding to different types of malignant tumors.

[0035] The data analysis module is used to perform data analysis and processing on the patient data matrix to obtain the weight value distribution of different malignant tumor factor characteristics within a preset time in the future and the weight score corresponding to the weight value distribution.

[0036] The model training module is used to create a prediction model and input the characteristics of malignant tumor factors, weight value distribution and weight score into the prediction model for training to obtain a preset model after training.

[0037] In the embodiments of the present application, lung cancer is used as an example, including but not limited to lung cancer as a malignant tumor. Lung cancer is a malignant tumor originating from the bronchial mucosa or glands of the lungs, with the fastest growing incidence and mortality rate, and one of the malignant tumors that poses the greatest threat to people's health and life. When conducting risk assessment on lung cancer, the malignant tumor factor characteristics that need to be collected include: age factor characteristics, smoking factor characteristics, chronic lung disease factor characteristics, occupational exposure factor characteristics, family history factor characteristics, etc., and then the patient data with lung cancer in the patient history data is determined for each malignant tumor. The factor characteristic data corresponding to the tumor factor characteristics, for example: age factor characteristics: 45 years old, smoking factor characteristics: 20 years (the number of packs smoked per day multiplied by the number of years of smoking), chronic lung disease factor: 1 (1 means suffering from chronic lung disease, 0 means not suffering from chronic lung disease), occupational exposure factor characteristics: 1 (1 means that the user has a certain history of occupational exposure, such as contact with asbestos, radon, beryllium, chromium, cadmium, nickel, silicon, coal smoke or coal smoke dust, etc., 0 means not having it), family history factor characteristics: 1 (1 means that there is a lung cancer patient among the first-degree relatives in the family, 0 means not having it).

[0038] The factor feature data are sorted according to the time series length to obtain the patient data matrix corresponding to different types of malignant tumors. The patient data matrix corresponding to each type of malignant tumor is: ; Among them, m is the characteristic of malignant tumor factors, and n is the length of the time series.

[0039] The patient data matrix is ​​then analyzed and processed to obtain the weight value distribution of different malignant tumor factor characteristics within a preset future time and the weight scores corresponding to the weight value distribution. A prediction model is then created, and the malignant tumor factor characteristics, weight value distribution, and weight scores are input into the prediction model for training to obtain a trained preset model.

[0040] Specifically, when performing data analysis and processing on the patient data matrix, it is determined whether the malignant tumor factor feature is an uncontrollable factor feature. If so, the data corresponding to the malignant tumor factor feature in the patient data matrix is ​​extracted and defined as the first sequence data, and the data corresponding to the remaining malignant tumor factor features in the patient data matrix is ​​defined as the second sequence data. Then, the data development corresponding to each malignant tumor factor feature in the first sequence data and the second sequence data is calculated according to the time series monitoring, and the free association coefficient corresponding to the first sequence data and the controllable association coefficient corresponding to the second sequence data are obtained. Then, based on the free association coefficient and the preset weight standard, each malignant tumor in the first sequence data is controlled. The weight values ​​of the factor characteristics in different unit times are calculated to obtain the first time weight group of different malignant tumor factor characteristics in the first sequence data, and then the weight values ​​of each malignant tumor factor characteristic in the second sequence data in different unit times are calculated based on the controllable correlation coefficient and the preset weight standard to obtain the second time weight group of different malignant tumor factor characteristics in the second sequence data, and the first time weight group and the second time weight group are respectively combined with the patient data matrix for data correspondence processing to generate a predicted data matrix, and then the predicted data matrix is ​​processed to obtain the weight value distribution of different malignant tumor factor characteristics in the future preset time and the weight score corresponding to the weight value distribution.

[0041] In the embodiment of the present application, the uncontrollable factor characteristics include: tumor deterioration caused by living environment factors, tumor deterioration caused by working environment factors, tumor deterioration caused by frequently traveled road factors, and tumor deterioration caused by age factors. Controllable factors include: tumor deterioration caused by malignant habit factors and tumor deterioration caused by physical fitness factors. The malignant tumor factor characteristics are classified according to the controllable factor characteristics and the uncontrollable factor characteristics, and are divided into first sequence data corresponding to the uncontrollable factor characteristics and second sequence data corresponding to the controllable factor characteristics, and then the correlation coefficient of the first sequence data and the second sequence data is calculated using the corresponding correlation coefficient calculation method.

[0042] Specifically, the method for calculating the free association coefficient of the first sequence data includes: retrieving the sequence data corresponding to different malignant tumor factor characteristics in the first sequence data, then performing numerical accumulation calculation on each sequence data, then performing mean calculation on the accumulated values ​​according to the interval span of the time series to obtain the mean corresponding to each sequence data, and performing cumulative mean calculation on all the means to obtain the free association coefficient corresponding to the first sequence data. The interval span of the time series represents the average time interval of each sequence data change in the first sequence data.

[0043] The calculation method of the controllable correlation coefficient of the second sequence data includes: calling the sequence data corresponding to different malignant tumor factor characteristics in the second sequence data, then performing numerical accumulation calculation on each sequence data, and then performing mean calculation on the accumulated value according to the controllable period segment and the uncontrollable period segment to obtain the mean corresponding to each sequence data, and performing the mean accumulation calculation on all the means to obtain the controllable correlation coefficient corresponding to the second sequence data. Among them, the controllable period segment indicates that each sequence data in the second sequence data has a time period segment that controls itself to reduce the occurrence of bad habits, and the uncontrollable period segment indicates that each sequence data in the second sequence data does not have a time period segment that controls itself to reduce the occurrence of bad habits.

[0044] For the embodiment of the present application, the preset weight calculation standard is the scoring standard and weight calculation standard corresponding to different pre-set free association coefficients and different controllable association coefficients. For example, when the free association coefficient is 30 and the controllable association coefficient is 70, the preset weight calculation standard for lung cancer is: Age factor feature: If the user is between 40 and 50 years old, 30 points will be assigned; if the age is greater than 50, 60 points will be assigned.

[0045] Smoking factor: If the user's smoking pack-years (number of packs smoked per day multiplied by the number of years of smoking) is greater than or equal to 20, 70 points will be assigned. Passive smoking factor: Check whether the user has lived or worked in the same room with a smoker for 20 years or more, if so, 30 points will be assigned.

[0046] Chronic lung disease factor: Users indicating that they have chronic lung diseases such as chronic obstructive pulmonary disease, tuberculosis, or pulmonary fibrosis will be assigned 30 points.

[0047] Occupational exposure factors: If the user has certain occupational exposure histories, such as exposure to asbestos, radon, beryllium, chromium, cadmium, nickel, silicon, coal smoke or coal smoke dust, 30 points will be assigned. Secondhand smoke or cooking fume factors: If the user selects exposure to secondhand smoke or cooking fume for 30 years or more, 30 points will be assigned.

[0048] Family history factor: If a user has a first-degree relative (parent, child, sibling) diagnosed with lung cancer, 30 points will be allocated.

[0049] The data values ​​in the patient data matrix are replaced with the corresponding score values, and the score weights are calculated according to the unit time in a row, where the unit time includes but is not limited to one year, and is specifically set according to the compactness of the data distribution in each row, and the proportion of the data with score values ​​in different years in each row to the total data in each row in that year is counted, so as to obtain the time weight value group of different malignant tumor factor characteristics. Then the weight value group is correspondingly integrated with the patient data matrix to generate an integrated data matrix, and the basic data distribution of the integrated data matrix is ​​explored to obtain the relative periodicity law of the weight value change of different malignant tumor factor characteristics in unit time, and the time period length is determined based on the relative periodicity law, and the integrated data matrix is ​​supervised based on the time period length. Time series data is sorted to obtain a sorted data matrix. Based on the sorted data matrix, the change trend of the weight value group in the future preset time period is predicted to generate prediction matrix data.

[0050] Specifically, according to the correspondence between the weight values ​​in the weight value group and the patient data matrix, the weight values ​​are integrated with the patient data matrix as dimensions. In the embodiment of the present application, pytorch technology is used for illustration, including but not limited to one implementable method of pytorch technology. The weight values ​​are added to the resource matrix data in a dimensional manner through the instruction "out.unsqueeze(-1)" in pytorch to achieve dimensional integration.

[0051] After integrating the weight value group with the patient data matrix, the following integrated data matrix is ​​obtained: ; Among them, v represents the weight value in the weight value group corresponding to the patient data matrix.

[0052] Specifically, the integrated data matrix is ​​imported into an Excel spreadsheet, and the Python integrated jupyter environment is configured to perform basic data distribution exploration on the n time series in the integrated data matrix. The main purpose is to find the relative periodicity of the corresponding sequence, and then determine the length of the time period based on the relative periodicity.

[0053] Specifically, t is used to replace the time periodicity length, and n in the integrated data matrix is ​​replaced by t to obtain a new sorted data matrix:

[0054] Specifically, assuming that the preset future time period is k, that is, the moving step length is k-step prediction, the prediction matrix data is obtained: ; Specifically, when the prediction data matrix is ​​processed to obtain the first training matrix, the normal distribution mean and the normal distribution variance of the feature data contained in the prediction data matrix are calculated, and the 3σ range of the prediction data matrix is ​​determined based on the normal distribution mean and the normal distribution variance, and then it is determined whether the feature data is outside the 3σ range. If the feature data is outside the 3σ range, the first matrix sequence of the prediction data matrix where the feature data is located is determined, the sequence average is calculated according to the first matrix sequence, and the feature data is replaced with the sequence average to obtain a second matrix sequence after replacement, and the second matrix sequence is missing value processed, and then the second matrix sequence in the prediction data matrix is ​​normalized to obtain the first training matrix.

[0055] Specifically, the 3σ range is based on the interference or noise of singular data caused by repeated measurements of equal precision under normal distribution, which is difficult to satisfy the normal distribution. If the absolute value of the residual error of a certain measurement value in a set of measurement data is νi>3σ, then the measurement value is a bad value and should be eliminated. Usually, the error equal to ±3σ is taken as the limit error. For the random error of normal distribution, the probability of falling outside ±3σ is only 0.27%. It is very unlikely to occur in a finite number of measurements, so there is a 3σ criterion. The 3σ criterion is the most commonly used and simplest criterion for judging gross errors. It is generally used when the number of measurements is sufficiently large (n≥30) or when n>10 is used for rough judgment.

[0056] Specifically, missing values ​​refer to data clustering, grouping, deletion or truncation caused by lack of information in a matrix sequence. The processing of missing values ​​is generally divided into deleting cases with missing values ​​and interpolating missing values. The embodiment of the present application uses the deletion of cases with missing values ​​to process the matrix sequence. There are mainly simple deletion and weighting methods for deleting cases with missing values. The simple deletion method is the most primitive method for processing missing values. It deletes cases with missing values. If the data missing problem can be achieved by simply deleting a small part of the sample, then this method is the most effective. When the type of missing value is non-completely randomly missing, the deviation can be reduced by weighting the complete data. After marking the cases with incomplete data, different weights are assigned to the complete data cases, and the weights of the cases can be obtained by logistic or probit regression.

[0057] Specifically, there are two forms of normalization methods: one is to convert the number into a decimal between (0, 1), and the other is to convert the dimensional expression into a dimensionless expression. It is mainly proposed for the convenience of data processing. Mapping the data to the range of 0 to 1 for processing is more convenient and faster.

[0058] The specific processing normalization method is: .

[0059] In an embodiment of the present application, when the first training matrix is ​​input into a preset algorithm model for data extrapolation to obtain the weight value distribution of different malignant tumor factor characteristics within a preset time in the future and the weight scores corresponding to the weight value distribution, a bidirectional LSTM model is used as an example of the preset algorithm model, including but not limited to a bidirectional LSTM model.

[0060] Specifically, the preset algorithm model is constructed, and the main body of the model adopts bidirectional LSTM as the trend prediction model. LSTM mainly consists of a forget gate, an input gate, and an output gate; Forget Gate: ; Input Gate: ; After the information is filtered by the forget gate and the input gate, the historical memory and the memory content of the current stage are merged, and the generated value is: ; Output Gate: ; After the LSTM described above, a layer of LSTM network layer is connected in reverse. Through this process, the BI-LSTM layer can be obtained, because several groups of weight combinations are trained together; add a spatial feature joint learning layer, initialize the association vector matrix size to M*V*K, take the output vector of the last layer of LSTM, transpose it and multiply it by the association vector parameter matrix, and finally connect the regression loss function to complete the construction of the preset algorithm model.

[0061] Specifically, the weight values ​​of different malignant tumor factor characteristics per unit time and the weight distribution corresponding to different malignant tumor factor characteristics in the future preset time period are determined according to the data output by the bidirectional LSTM model. Then a prediction model is created, and the malignant tumor factor characteristics, the weight distribution corresponding to the malignant tumor factor characteristics, and the weight score are input into the prediction model for training to obtain the final trained preset model, wherein the preset model is a neural network model.

[0062] In the embodiment of the present application, the weight value represents the scoring standard of different malignant tumor factor characteristics per unit time. For example, the weight value of the starting age factor characteristic is 30%: if the user is between 40 and 50 years old, 30 points will be allocated; if the age is greater than 50, 50 points will be allocated. When the weight value of the age factor characteristic obtained after the predicted calculation is 25%, if the user is between 40 and 50 years old, 25 points will be allocated, and if the age is greater than 50 years old, 42 points will be allocated.

[0063] When the user's data parameter characteristics and parameter characteristic values ​​are input into the prediction model for training, the data parameter characteristics will first be matched with the malignant tumor factor characteristics in the prediction model to obtain the weight distribution and weight score corresponding to the data parameter characteristics. Then, the characteristics with evaluable scores in the parameter characteristics will be determined based on the parameter characteristic values, and the disease probability values ​​corresponding to the characteristics at different time stages will be determined based on the weight distribution and weight score.

[0064] The diagnosis feedback module is used to convert the format of the disease probability value based on a preset conversion standard, obtain the disease probability information, and feed back the disease probability information to the target device.

[0065] Specifically, the preset conversion standard is the disease probability information corresponding to different disease probability values. The disease probability value of the current user is matched with the probability value in the preset conversion standard to obtain the disease probability information of the user. For example, the disease probability information includes: Your disease probability value is 72, 72 is a high-risk probability value, and it is recommended that you **.

[0066] For the embodiments of the present application, breast cancer is taken as an example to illustrate the risk diagnosis decision. When making diagnostic decisions for breast cancer patients, low-risk people are still recommended to carry out breast cancer screening after the age of 50, and the screening content refers to the selection of medium-risk people. Medium-risk people should undergo regular breast cancer-related screening. Start breast ultrasound examination at the age of 40. Start breast ultrasound and breast mammography at the age of 50. The interval is one year. High-risk people need to start breast cancer screening at any age diagnosed as high risk, mainly breast ultrasound. In addition, breast ultrasound and mammography screening are required at the age of 45. First-degree blood relatives who are diagnosed with breast cancer or ovarian cancer before the age of 50 are recommended to undergo genetic risk testing (BRCA-related gene testing); extremely high-risk people with scores above 80 or clear genetic risk carriers need to undergo baseline breast MRI examinations at the first screening.

[0067] Follow-up and treatment of imaging screening Breast ultrasound, X-ray and MRI examinations should all be recorded and analyzed according to the Breast Imaging Reporting and Data System (BI-RADS). For example: 1. BI-RADS Category 1 and BI-RADS Category 2: Regular screening, no special treatment required.

[0068] 2. BI-RADS Category 3: (1) X-ray assessment of Category 3 lesions: It is recommended to perform a mammographic review of the affected breast after 6 months, and a bilateral breast review at 12 and 24 months. If the lesion remains stable, follow-up can be continued; if there is no change during 2 to 3 years of follow-up, it can be downgraded to BI-RADS Category 2; if the lesion shrinks or disappears during follow-up, it can be downgraded to BI-RADS Category 2 or BI-RADS Category 1; if there are suspicious changes in the lesion during follow-up, a biopsy should be considered to clarify the pathological nature. (2) Ultrasound assessment of BI-RADS Category 3 lesions: It is recommended to perform an ultrasound follow-up review after 3 to 6 months. If there is no change during 2 years of follow-up, it can be downgraded to BI-RADS Category 2; if there are suspicious changes in the lesion during follow-up, a biopsy should be considered to clarify the pathological nature.

[0069] 3. BI-RADS 4A: Closely observe changes in the lesion and perform a biopsy to clarify the pathological nature if necessary.

[0070] 4. BI-RADS 4B: Further imaging examination and biopsy can be performed.

[0071] 5. BI-RADS 4C and 5 categories: The evaluation is incomplete and requires comparison with previous examinations and recommendation for re-examination or other imaging examinations for comprehensive evaluation.

[0072] 6 BI-RADS Category 0: The assessment is incomplete and requires comparison with previous examinations and recommendation for reexamination or other imaging examinations for comprehensive evaluation.

[0073] In an embodiment of the present application, when a user is being assessed for malignant tumors, personal information of the user input into a target device is obtained, and then the user data in the user information is processed to obtain user verification data, and data features are extracted from the user verification data to obtain data parameter features corresponding to the user and parameter feature values ​​corresponding to the data parameter features. The data parameter features and parameter feature values ​​are then input into a prediction model for training to obtain disease probability values ​​of the user at different time stages, and then the disease probability values ​​are formatted based on a preset conversion standard to obtain disease probability information, and the disease probability information is fed back to the target device. By viewing the information fed back by the target device, the user can know the disease probability values ​​corresponding to the user at different time stages, thereby achieving the effect of improving the accuracy of malignant tumor risk assessment.

[0074] A possible implementation of the embodiment of the present application is to perform data analysis on the patient data matrix to obtain the weight value distribution of different malignant tumor factor characteristics within a preset time in the future and the weight score corresponding to the weight value distribution, and then further include: Determine the actual weight distribution and actual weight score detected for different malignant tumor factor characteristics within a preset time period in the future, and then judge whether the actual weight distribution and actual weight score are consistent with the weight value distribution and weight score. If not, denormalize the weight distribution and weight score based on the actual weight distribution and actual weight score, and restore the weight distribution and weight score to the actual weight distribution and actual weight score.

[0075] A possible implementation of the embodiment of the present application is to feed back the disease probability information to the target device, and then further include: Based on the disease probability information, the user's disease characteristics and the risk level corresponding to the disease characteristics are determined, and then the disease characteristics and risk level are input into the trained prevention model to obtain prevention suggestion information, and the prevention suggestion information is sent to the target device.

[0076] An electronic device is provided in an embodiment of the present application, such as Figure 3 As shown, Figure 3 The electronic device 300 shown includes: a processor 301 and a memory 303. The processor 301 and the memory 303 are connected, such as through a bus 302. Optionally, the electronic device 300 may also include a transceiver 304. It should be noted that in actual applications, the transceiver 304 is not limited to one, and the structure of the electronic device 300 does not constitute a limitation on the embodiments of the present application.

[0077] The processor 301 may be a CPU (Central Processing Unit), a general-purpose processor, a DSP (Digital Signal Processor), an ASIC (Application Specific Integrated Circuit), an FPGA (Field Programmable Gate Array) or other programmable logic devices, transistor logic devices, hardware components or any combination thereof. It may implement or execute various exemplary logic blocks, modules and circuits described in conjunction with the disclosure of this application. The processor 301 may also be a combination that implements computing functions, such as a combination of one or more microprocessors, a combination of a DSP and a microprocessor, etc.

[0078] The bus 302 may include a path to transmit information between the above components. The bus 302 may be a PCI (Peripheral Component Interconnect) bus or an EISA (Extended Industry Standard Architecture) bus. The bus 302 may be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Figure 3 Only one thick line is used in the diagram, but this does not mean that there is only one bus or only one type of bus.

[0079] The memory 303 may be a ROM (Read Only Memory) or other types of static storage devices that can store static information and instructions, a RAM (Random Access Memory) or other types of dynamic storage devices that can store information and instructions, or an EEPROM (Electrically Erasable Programmable Read Only Memory), a CD-ROM (Compact Disc Read Only Memory) or other optical disk storage, optical disk storage (including compressed optical disk, laser disk, optical disk, digital versatile disk, Blu-ray disk, etc.), a magnetic disk storage medium or other magnetic storage device, or any other medium that can be used to carry or store the desired program code in the form of instructions or data structures and can be accessed by a computer, but is not limited thereto.

[0080] The memory 303 is used to store the application code for executing the solution of the present application, and the execution is controlled by the processor 301. The processor 301 is used to execute the application code stored in the memory 303 to implement the contents shown in the above system embodiment.

[0081] The electronic devices include, but are not limited to, mobile terminals such as mobile phones, laptop computers, digital broadcast receivers, PDAs (personal digital assistants), PADs (tablet computers), PMPs (portable multimedia players), vehicle-mounted terminals (such as vehicle-mounted navigation terminals), and fixed terminals such as digital TVs, desktop computers, etc. It can also be a server, etc. Figure 3 The electronic device shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present disclosure.

[0082] The embodiment of the present application provides a computer-readable storage medium, on which a computer program is stored, and when the computer is run on a computer, the computer can execute the corresponding content in the above-mentioned system embodiment. Compared with the related art, in the embodiment of the present application, by adopting the above-mentioned technical solution, when the user is evaluated for malignant tumors, the user's personal information input to the target device is obtained, and then the user data in the user information is processed to obtain user verification data, and the user verification data is subjected to data feature extraction to obtain data parameter features corresponding to the user and parameter feature values ​​corresponding to the data parameter features, and then the data parameter features and parameter feature values ​​are input into the prediction model for training to obtain the user's disease probability value at different time stages, and then the disease probability value is formatted based on the preset conversion standard to obtain disease probability information, and the disease probability information is fed back to the target device, and the user can know the disease probability value corresponding to the user at different time stages by viewing the information fed back by the target device, thereby achieving the effect of improving the accuracy of malignant tumor risk assessment.

[0083] It should be understood that, although the steps in the flowchart of the accompanying drawings are displayed in sequence as indicated by the arrows, these steps are not necessarily executed in sequence in the order indicated by the arrows. Unless otherwise specified herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least a part of the steps in the flowchart of the accompanying drawings may include multiple sub-steps or multiple stages, and these sub-steps or stages are not necessarily executed at the same time, but can be executed at different times, and their execution order is not necessarily sequential, but can be executed in turn or alternately with other steps or at least a part of the sub-steps or stages of other steps.

[0084] The above are only some implementation methods of the present application. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present application. These improvements and modifications should also be regarded as the scope of protection of the present application.

Claims

1. A malignant tumor risk assessment and diagnosis decision system, characterized in that: include: An information acquisition module, used to acquire user information, wherein the user information is personal information input by the user through the target device; A data processing module, used to process the user data in the user information to obtain user verification data; A feature extraction module, used to extract data features from the user verification data to obtain data parameter features corresponding to the user and parameter feature values ​​corresponding to the data parameter features; A feature evaluation module, used for inputting the data parameter features and the parameter feature values ​​into a prediction model for training, and obtaining the disease probability values ​​of the user at different time stages; The diagnosis feedback module is used to convert the format of the disease probability value based on a preset conversion standard to obtain disease probability information, and feed back the disease probability information to the target device.

2. A malignant tumor risk assessment and diagnosis decision system according to claim 1, characterized in that: The system also includes: a feature data acquisition module, a data determination module, a data sorting module, a data analysis module and a model training module, wherein: The characteristic data acquisition module is used to acquire the characteristics of malignant tumor factors corresponding to different types of malignant tumors and the historical data of patients suffering from different types of malignant tumors; The data determination module is used to determine the factor characteristic data corresponding to each malignant tumor factor characteristic and the time series length corresponding to the factor characteristic data according to the patient historical data; The data sorting module is used to perform unsupervised time series data sorting on the patient historical data according to the factor feature data and the time series length to obtain patient data matrices corresponding to different types of malignant tumors; The data analysis module is used to perform data analysis on the patient data matrix to obtain the weight value distribution of different malignant tumor factor characteristics within a preset time in the future and the weight scores corresponding to the weight value distribution; The model training module is used to create a prediction model, and input the malignant tumor factor characteristics, weight value distribution and weight score into the prediction model for training to obtain a trained prediction model.

3. A malignant tumor risk assessment and diagnosis decision system according to claim 2, characterized in that: When the data analysis module performs data analysis on the patient data matrix to obtain the weight value distribution of different malignant tumor factor characteristics within a preset time in the future and the weight score corresponding to the weight value distribution, it is specifically used to: Determine whether the malignant tumor factor feature is an uncontrollable factor feature, and if so, extract and define the data corresponding to the malignant tumor factor feature in the patient data matrix as the first sequence data, and define the remaining data corresponding to the malignant tumor factor feature in the patient data matrix as the second sequence data; Calculate the data development corresponding to each malignant tumor factor feature in the first sequence data and the second sequence data according to the time series monitoring, and obtain the free association coefficient corresponding to the first sequence data and the controllable association coefficient corresponding to the second sequence data; Calculating the weight value of each malignant tumor factor feature in the first sequence data in different unit times based on the free association coefficient and the preset weight standard to obtain a first time weight group of different malignant tumor factor features in the first sequence data; Calculating the weight value of each malignant tumor factor feature in the second sequence data in different unit times based on the controllable correlation coefficient and the preset weight standard to obtain a second time weight group of different malignant tumor factor features in the second sequence data; Respectively performing data correspondence and combining processing on the first time weight group and the second time weight group with the patient data matrix to generate a prediction data matrix; The prediction data matrix is ​​processed to obtain the weight value distribution of different malignant tumor factor characteristics within a preset future time and the weight scores corresponding to the weight value distribution.

4. A malignant tumor risk assessment and diagnosis decision system according to claim 3, characterized in that: When the data analysis module processes the prediction data matrix to obtain the weight value distribution of different malignant tumor factor characteristics within a preset time in the future and the weight score corresponding to the weight value distribution, it is specifically used to: Performing data processing on the prediction data matrix to obtain a first training matrix; The first training matrix is ​​input into a preset algorithm model for data extrapolation to obtain the weight value distribution of different malignant tumor factor characteristics within a future preset time and the weight score corresponding to the weight value distribution.

5. A malignant tumor risk assessment and diagnosis decision system according to claim 3, characterized in that: When the data analysis module performs data corresponding and combining processing on the first time weight group and the second time weight group with the patient data matrix to generate a prediction data matrix, the data analysis module is specifically used to: respectively integrating the first time weight group and the second time weight group with the patient data matrix to generate an integrated data matrix; Performing basic data distribution exploration on the integrated data matrix to obtain the relative periodicity of the weight value changes of the different malignant tumor factor characteristics within a unit time, and determining the time period length based on the relative periodicity; Performing supervised time series data sorting on the integrated data matrix based on the time period length to obtain a sorted data matrix; Based on the sorted data matrix, the change trend of the weight value group within a future preset time period is predicted to generate prediction matrix data.

6. A malignant tumor risk assessment and diagnosis decision system according to claim 4, characterized in that: When the data analysis module processes the prediction data matrix to obtain the first training matrix, it is specifically used to: Calculating a normal distribution mean and a normal distribution variance of the feature data contained in the prediction data matrix, and determining a 3σ range of the prediction data matrix based on the normal distribution mean and the normal distribution variance; Determine whether the feature data is outside the 3σ range, if the feature data is outside the 3σ range, determine a first matrix sequence of the prediction data matrix where the feature data is located, calculate a sequence average value according to the first matrix sequence, replace the feature data with the sequence average value, obtain a replaced second matrix sequence, and perform missing value processing on the second matrix sequence; A sequence normalization process is performed on the second matrix sequence in the prediction data matrix to obtain a first training matrix.

7. A malignant tumor risk assessment and diagnosis decision system according to claim 4, characterized in that: The system further comprises: a weight determination module and a weight reset module, wherein: The weight determination module is used to determine the actual weight distribution and actual weight score of the different malignant tumor factor characteristics detected within a future preset time period; The weight resetting module is used to determine whether the actual weight distribution and the actual weight score are consistent with the weight value distribution and the weight score respectively. If they are inconsistent, the weight distribution and the weight score are denormalized based on the actual weight distribution and the actual weight score to restore the weight distribution and the weight score to the actual weight distribution and the actual weight score.

8. The malignant tumor risk assessment and diagnosis decision system according to claim 1, characterized in that: The system further comprises: a disease determination module and a suggestion generation module, wherein: The disease determination module is used to determine the disease characteristics of the user and the risk level corresponding to the disease characteristics based on the disease probability information; The suggestion generation module is used to input the disease characteristics and risk levels into the trained prevention model to obtain prevention suggestion information, and send the prevention suggestion information to the target device.

9. An electronic device, characterized in that: The electronic device includes: one or more processors; Memory; One or more applications, wherein the one or more applications are stored in a memory and configured to be executed by one or more processors, and the one or more applications are configured to: execute a malignant tumor risk assessment and diagnosis decision system as described in any one of claims 1 to 8.

10. A computer-readable storage medium, characterized in that: include: The computer program of a malignant tumor risk assessment and diagnosis decision system according to any one of claims 1 to 8 is stored and can be loaded by a processor and executed.

Images