Gene classification method, system and device based on multi-label feature selection

Abstract

The present invention provides a gene classification method, system and device based on multi-label feature selection, the method comprising: obtaining a gene expression data set, performing feature selection on the gene expression data set using a multi-label feature selection model based on missing gene data, and obtaining a feature subset of the gene expression data set; training a gene classification model using the feature subset of the gene expression data set, and obtaining a trained gene classification model; classifying the genes to be classified based on the feature subset of the gene data to be classified using the trained gene classification model, and obtaining labels of the gene data to be classified. The present invention adopts a gene classification method based on multi-label feature selection, completes feature selection of multi-label gene data based on missing data, considers all dependencies between gene labels and features, and completes gene classification based on the feature subset after feature selection, thereby fundamentally solving the problem of low accuracy of existing multi-label gene classification.

Description

Technical Field

[0001] The present invention belongs to the field of bioinformatics, and specifically relates to a gene classification method, system and equipment based on multi-label feature selection. Background Art

[0002] With the development of DNA microarray technology, a huge amount of gene expression data has been generated. Gene expression data contains rich information about gene activities. Analyzing the patterns hidden in gene expression data is of great significance for understanding and inferring biological gene functions and studying gene regulation mechanisms. How to effectively analyze massive amounts of gene expression data has become an important challenge in the field of bioinformatics.

[0003] At present, gene classification is a very popular research field. Gene expression data is generally in the form of a matrix, with the characteristics of high dimension, small sample and multiple labels. The main feature of multi-label gene data is that a sample can be associated with multiple labels at the same time. Multi-label data contains three kinds of variable relationships, namely the correlation between labels, between features and between labels and between features. High-dimensional gene data with a large number of redundant features significantly increases the computational burden of multi-label gene data classification, and also leads to overfitting and performance degradation of gene classification, which greatly reduces the accuracy of gene classification results. At the same time, the existing multi-label gene feature selection methods do not pay enough attention to the role of gene labels, ignore the relationship within the label, and rarely reveal its potential causal mechanism. Instead, they study the correlation between features and labels, the correlation between labels and labels, or the correlation between features and features separately. They rarely handle the three correlations at the same time and ignore their mutual influence. At present, there is a problem that the accuracy of gene classification is still low after the above multi-label feature selection method.

[0004] Therefore, how to perform multi-label feature selection of genetic data and improve the accuracy of gene classification is an urgent problem to be solved in this field. Summary of the invention

[0005] The purpose of the present invention is to provide a gene classification method, system and device based on multi-label feature selection in view of the defects of the prior art. The gene classification method based on multi-label feature selection of the present invention completes the feature selection of multi-label gene data based on missing data, incorporates the causal relationship between gene labels, and alternately uses the mutually reinforcing missing value interpolation process and multi-label MB learning process. At the same time, the method of the present invention takes into account all the dependencies between gene labels and gene features, and completes gene classification based on the feature subset after feature selection, thereby solving the problem of low accuracy of existing multi-label gene classification from the root.

[0006] In order to achieve the above purpose, the present invention adopts the following technical solutions:

[0007] The present invention proposes a gene classification method based on multi-label feature selection, which is characterized by comprising the steps of:

[0008] S1. Construct a multi-label feature selection model based on missing gene data;

[0009] S2, obtaining a gene expression data set, and performing feature selection on the gene expression data set using the multi-label feature selection model based on missing gene data to obtain a feature subset of the gene expression data set;

[0010] Step S2 specifically includes the following steps:

[0011] S21, obtaining a gene expression dataset, wherein the gene expression dataset contains m-dimensional features and q-dimensional labels;

[0012] S22, genetic data interpolation;

[0013] S23, find the multi-label Markov blanket MB, and use the multi-label causal feature selection learning method MLMB to learn the MB of the class variable;

[0014] S24, obtaining an extended-MB; the extended-MB is composed of the union of the MB of the class variable and the MB of each variable in the class variable MB;

[0015] S25, updating the data set; updating the data set to a new data set containing only extended-MB;

[0016] S26, determine whether the iteration termination condition is met, if so, stop, if not, repeat steps S22-S25 continuously;

[0017] S27, the iteration ends and the feature subset of the gene expression data set is returned;

[0018] S3, training a gene classification model using a feature subset of the gene expression data set to obtain a trained gene classification model;

[0019] S4. Obtain gene data to be classified, and classify the genes to be classified using a trained gene classification model based on the feature subset of the gene data to be classified to obtain labels for the gene data to be classified.

[0020] Furthermore, in the gene expression dataset, F = {F1, F2, ..., F m} represents an m-dimensional feature set, Y = {Y1, Y2, ..., Y q} represents a q-dimensional label set, V = F ∪ {Y}, S is any set of variables in V; (where i≠j, Indicates V i Conditionally independent of V under given s j ; V\Y i With V\{Y i} is equivalent to, which means that in V, Y is divided i All features except

[0021] Furthermore, the gene data interpolation described in step S22 specifically includes: using the observation information of the incomplete instances and the complete instances in the gene expression data set to estimate the missing values, and using KNN or Lagrange interpolation method to interpolate the gene data.

[0022] Further, step S23 includes:

[0023] (1) Mining the causal mechanism of each gene signature; from all features and all except Y i Learn each Y from the label i MB of class labels, i.e. from V\Y i Find Y i MB, and record the result as MB(Y i );

[0024] (2) Detect features that are ignored due to label correlation; specifically, from F\MB(Y i ) to detect features that are ignored due to strong label correlation;

[0025] (3) Correct the false features obtained; specifically, according to MB(Y i ) and Y i The correlation degree is arranged in ascending order, and the top k2% features with the weakest correlation are selected and stored in SelFea(Y i ) in; traverse q tags and detect SelFea(Y i ) whether the MB of the feature contains Y i , from SelFea(Y i ) to remove false MB features.

[0026] Furthermore, the gene classification model adopts MLKNN.

[0027] The present invention also proposes a gene classification system based on multi-label feature selection, characterized in that the gene classification system executes the gene classification method based on multi-label feature selection, including: a multi-label feature selection model construction module, a feature selection module, a gene classification model training module, and a gene classification module;

[0028] Multi-label feature selection model construction module, which constructs a multi-label feature selection model based on missing gene data;

[0029] A feature selection module is used to obtain a gene expression data set, and perform feature selection on the gene expression data set using the multi-label feature selection model based on missing gene data to obtain a feature subset of the gene expression data set;

[0030] A gene classification model training module is used to train a gene classification model using a feature subset of the gene expression data set to obtain a trained gene classification model;

[0031] The gene classification module obtains the gene data to be classified, and based on the feature subset of the gene data to be classified, classifies the gene to be classified using a trained gene classification model to obtain a label of the gene data to be classified.

[0032] The present invention further provides a computer device, comprising a memory and a processor, wherein a computer program is stored in the memory, and the processor implements the above method when executing the computer program.

[0033] Compared with the prior art, the present invention has the following beneficial effects: the gene classification method based on multi-label feature selection of the present invention first completes the feature selection of multi-label gene data based on missing data, incorporates the causal relationship between gene labels, and utilizes the mutually reinforcing missing value interpolation process and multi-label MB learning process alternately. At the same time, the method of the present invention takes into account all the dependencies between gene labels and gene features; and then completes gene classification based on the feature subset after feature selection, thereby improving the accuracy of gene classification and reducing the computational burden of high-dimensional gene classification. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying any creative work.

[0035] Figure 1 A schematic diagram of a gene classification method based on multi-label feature selection provided in an embodiment of the present invention.

[0036] Figure 2 A schematic diagram of a multi-label feature selection model based on missing gene data provided in an embodiment of the present invention.

[0037] Figure 3 A schematic diagram of a process for obtaining a multi-label MB according to an embodiment of the present invention.

[0038] Figure 4 Schematic diagram of a gene classification system based on multi-label feature selection provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0039] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0040] It should be understood that when used in this specification and the appended claims, the terms "include" and "comprises" indicate the presence of described features, integers, steps, operations, elements and / or components, but do not exclude the presence or addition of one or more other features, integers, steps, operations, elements, components and / or combinations thereof.

[0041] The present invention will be further described below in conjunction with the accompanying drawings and specific embodiments, but they are not intended to limit the present invention.

[0042] It should also be understood that the terms used in this specification of the present invention are only for the purpose of describing specific embodiments and are not intended to limit the present invention. As used in the specification of the present invention and the appended claims, unless the context clearly indicates otherwise, the singular forms "a", "an" and "the" are intended to include plural forms.

[0043] It should be further understood that the term "and / or" used in the present description and the appended claims refers to any and all possible combinations of one or more of the associated listed items, and includes these combinations.

[0044] The present invention provides a gene classification method based on multi-label feature selection. Figure 1 As shown, the gene classification method based on multi-label feature selection includes the following steps S1 to S4.

[0045] S1. Construct a multi-label feature selection model based on missing gene data;

[0046] S2, obtaining a gene expression data set, and performing feature selection on the gene expression data set using the multi-label feature selection model based on missing gene data to obtain a feature subset of the gene expression data set; the process of step S2 is as follows Figure 2 As shown;

[0047] Step S2 specifically includes the following steps:

[0048] S21, obtaining a gene expression dataset, wherein the gene expression dataset contains m-dimensional features and q-dimensional labels;

[0049] In the gene expression dataset, F = {F1, F2, ..., F m} represents an m-dimensional feature set, Y = {Y1, Y2, ..., Y q} represents a q-dimensional label set, V = F ∪ {Y}, s is any set of variables in V; (where i≠j, Indicates V i Conditionally independent of V given S j ; V\Y i With V\{Y i} is equivalent to, which means that in V, Y is divided i All features except

[0050] S22, genetic data interpolation;

[0051] Specifically, data imputation uses the observation information of incomplete instances and complete instances in the dataset to estimate missing values, so it enables multi-label causal learning methods to address most of the missing values ​​in the dataset, especially when there are more incomplete instances than complete instances. Data imputation provides an accurate dataset for the reliability of MB learning.

[0052] In a specific embodiment, KNN or Lagrange interpolation method is used to interpolate genetic data.

[0053] S23, find the multi-label Markov blanket MB, and use the multi-label causal feature selection learning method MLMB to learn the class variable MB; the specific process of step S23 is as follows Figure 3 As shown;

[0054] Specifically, the MB of the learning class variable is found by a perfect multi-label causal feature selection algorithm MLMB.

[0055] Step S23 includes:

[0056] (1) Mining the causal mechanism of each gene signature; from all features and all except Y i Learn each Y from the label i MB of class labels, i.e. from V\Y i Find Y i MB, and record the result as MB(Y i );

[0057] (2) Detect features that are ignored due to label correlation; specifically, from F\MB(Y i ) to detect features that are ignored due to strong label correlation.

[0058] If we directly detect each ignored feature, it will be computationally very expensive. To solve the high-dimensional problem, we cani The correlation degree is sorted in descending order, and the top k1% features with the highest correlation are selected and stored in sel(Y i ) to detect, selectively from sel(Y i ) to restore the lost true MB features.

[0059] (3) Correct the false features obtained; specifically, according to MB(Y i ) in ascending order of the degree of association between the features and , select MB(Y i ) and Y i The weakest correlation of the first k2% features is stored in SelFea(Y i ), traverse q tags, and detect SelFea(Y i ) whether the MB of the feature contains Y i , from SelFea(Y i ) to remove false MB features.

[0060] For label Y i ∈Y, in MB(Y i ) always contains a small number of erroneous features. So according to MB(Y i ) and Y i The degree of association is in ascending order, select MB(Y i ) and Y i The top k2% features with the weakest correlation are stored in SalFea(Y i ) for testing.

[0061] S24, obtaining an extended-MB; the extended-MB is composed of the union of the MB of the class variable and the MB of each variable in the class variable MB.

[0062] The rationale behind this strategy is that, according to the MB feature selection theory, since missing values ​​are filled with reasonable values ​​using data interpolation methods, for a noisy dataset, the extended-MB of the class variable has a higher probability of containing causal informative features than the MB of the class variable.

[0063] S25, updating the data set; updating the data set to a new data set containing only extended-MB.

[0064] Specifically, the new extended-MB data set found in step S24 enables data interpolation to fill in the missing values ​​of only the causal information features, rather than all the features in the data set.

[0065] S26, determine whether the iteration termination condition is met, if so, stop, if not, repeat steps S22-S25 continuously;

[0066] In a specific embodiment, the iteration termination condition is that the data in the extended-MB no longer changes.

[0067] S27. After the iteration, the feature subset of the gene expression data set is returned.

[0068] In a specific embodiment, the returned optimal feature subset is used to test the classification accuracy of the test set, and the quality of the feature subset of the gene expression data set obtained is verified by the obtained classification accuracy.

[0069] S3. Using the feature subset of the gene expression data set, the gene classification model is trained to obtain a trained gene classification model.

[0070] The gene classification model adopts MLKNN. The feature subset of the gene expression data set is input into the MLKNN model for training, the number of parameters k of the MLKNN model is 10, and other parameters remain default, to obtain the MLKNN model after the data set is optimized.

[0071] In bioinformatics research, yeast gene expression data is often used to verify the application effects of theoretical models and algorithms.

[0072] In a specific embodiment, the gene expression data set: the yeast data set is a typical multi-label gene expression data set, which contains the microarray expression and phylogenetic map of 2417 yeast genes. Each generation is annotated with a subset of the 14 functional categories (such as Metabolism, energy, etc.) at the top level of the functional catalog. In order to test the performance of the method proposed in this application, in addition to the class attributes in the data set, four different levels of missing values ​​are set in each feature: 5%, 10%, 15% and 20%, to generate a data set with missing values.

[0073] We use Hamming Loss, Average Precision, Coverage, and Ranking Loss as evaluation criteria for classification models:

[0074] 1) Hamming Loss:

[0075]

[0076] 2)Average Precision:

[0077]

[0078] 3) Coverage:

[0079]

[0080] 4)Ranking Loss:

[0081]

[0082] According to the steps of the multi-label feature selection process diagram of the overall flow chart of the invention, the feature subset MB we need is returned after the process is completed, and finally the feature subset MB is used to train the MLKNN classifier model to obtain the model MLKNN_MB. For comparative experiments, we use the original data without feature selection and directly train the MLKNN model with the training set Train to obtain the model MKLNN_Train. Substitute the test set Test to obtain the four indicators of the MLKNN_Train model. The above data are aggregated into the following Table 1:

[0083] Table 1: Comparison of four indicators between feature subset MB and all feature dataset scene

[0084]

[0085] The Average Precision index in the table is as large as possible, while the Hamming Loss, Coverage and Ranking Loss indexes are as small as possible. From the above experimental results, it can be seen that the MLKNN_MB classifier is better than the MLKNN_Train classifier in many indicators. This shows that the gene classification method based on multi-label feature selection proposed by us can effectively improve the classification accuracy.

[0086] S4. Obtain gene data to be classified, and classify the genes to be classified using a trained gene classification model based on the feature subset of the gene data to be classified to obtain labels for the gene data to be classified.

[0087] The present invention proposes a gene classification method based on multi-label feature selection. First, the data is initialized, and the causal structure of each class label is learned. The multi-label missing data is selected using multiple interpolation MBs. First, the multi-label MBs are found, and then the extended-MBs are obtained. Data interpolation and multi-label MB learning are integrated in a unified framework so that the two modules can cooperate with each other. MB learning helps to interpolate missing data in potential causal features, and data interpolation provides an accurate interpolation data set for the reliability of MB learning. Gene classification is then completed based on the feature subset after feature selection, which improves the accuracy of gene classification and reduces the computational burden of high-dimensional gene classification.

[0088] Figure 4 This is a gene classification system based on multi-label feature selection provided by an embodiment of the present invention. Figure 4 As shown, the gene classification system based on multi-label feature selection includes a multi-label feature selection model construction module, a feature selection module, a gene classification model training module, and a gene classification module;

[0089] Multi-label feature selection model construction module, which constructs a multi-label feature selection model based on missing gene data;

[0090] A feature selection module is used to obtain a gene expression data set, and perform feature selection on the gene expression data set using the multi-label feature selection model based on missing gene data to obtain a feature subset of the gene expression data set;

[0091] A gene classification model training module is used to train a gene classification model using a feature subset of the gene expression data set to obtain a trained gene classification model;

[0092] The gene classification module obtains the gene data to be classified, and based on the feature subset of the gene data to be classified, classifies the gene to be classified using a trained gene classification model to obtain a label of the gene data to be classified.

[0093] The above-mentioned gene classification system based on multi-label feature selection can be implemented in the form of a computer program, which can be run on a computer device.

[0094] The computer device may be a server, wherein the server may be an independent server or a server cluster composed of multiple servers.

[0095] The computer device includes a processor, a memory, and a network interface connected via a system bus, wherein the memory may include a non-volatile storage medium and an internal memory.

[0096] The non-volatile storage medium can store an operating system and a computer program. The computer program includes program instructions, and when the program instructions are executed, the processor can execute a gene classification method based on multi-label feature selection.

[0097] The processor is used to provide computing and control capabilities to support the operation of the entire computer equipment.

[0098] The internal memory provides an environment for the operation of a computer program in a non-volatile storage medium. When the computer program is executed by a processor, the processor can execute a gene classification method based on multi-label feature selection.

[0099] The network interface is used to communicate with other devices over the network. Those skilled in the art will appreciate that the above-mentioned computer device structure is only a partial structure related to the present application solution, and does not constitute a limitation on the computer device to which the present application solution is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0100] The processor is used to run a computer program stored in the memory, and the program implements the gene classification method based on multi-label feature selection described in Example 1.

[0101] It should be understood that in the embodiments of the present application, the processor may be a central processing unit (CPU), and the processor may also be other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. Among them, the general-purpose processor may be a microprocessor or the processor may also be any conventional processor, etc.

[0102] It can be understood by those skilled in the art that all or part of the processes in the method for implementing the above embodiment can be completed by instructing the relevant hardware through a computer program. The computer program includes program instructions, and the computer program can be stored in a storage medium, which is a computer-readable storage medium. The program instructions are executed by at least one processor in the computer system to implement the process steps of the embodiment of the above method.

[0103] The present invention also provides a storage medium. The storage medium may be a computer-readable storage medium. The storage medium stores a computer program, wherein when the computer program is executed by a processor, the processor executes a gene classification method based on multi-label feature selection as described in Embodiment 1.

[0104] The storage medium may be a USB flash drive, a mobile hard disk, a read-only memory (ROM), a magnetic disk, or an optical disk, or any other computer-readable storage medium that can store program codes.

[0105] Those of ordinary skill in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of the two. In order to clearly illustrate the interchangeability of hardware and software, the composition and steps of each example have been generally described in terms of function in the above description. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of the present invention.

[0106] In the several embodiments provided by the present invention, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of each unit is only a logical function division, and there may be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed.

[0107] The steps in the method of the embodiment of the present invention can be adjusted in order, combined and deleted according to actual needs. The units in the device of the embodiment of the present invention can be combined, divided and deleted according to actual needs. In addition, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0108] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a storage medium. Based on this understanding, the technical solution of the present invention is essentially or the part that contributes to the prior art, or all or part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions for a computer device (which can be a personal computer, terminal, or network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention.

[0109] Note that the above are only preferred embodiments of the present invention and the technical principles used. Those skilled in the art will understand that the present invention is not limited to the specific embodiments described herein, and that various obvious changes, readjustments and substitutions can be made by those skilled in the art without departing from the scope of protection of the present invention. Therefore, although the present invention has been described in more detail through the above embodiments, the present invention is not limited to the above embodiments, and may include more other equivalent embodiments without departing from the concept of the present invention, and the scope of the present invention is determined by the scope of the appended claims.

Claims

1. A gene classification method based on multi-label feature selection, characterized in that: Includes steps: S1. Construct a multi-label feature selection model based on missing gene data; S2, obtaining a gene expression data set, and performing feature selection on the gene expression data set using the multi-label feature selection model based on missing gene data to obtain a feature subset of the gene expression data set; Step S2 specifically includes the following steps: S21, obtaining a gene expression dataset, wherein the gene expression dataset contains m-dimensional features and q-dimensional labels; S22, genetic data interpolation; The gene data interpolation described in step S22 specifically includes: using the observation information of the incomplete instances and the complete instances in the gene expression data set to estimate the missing values, and using the Lagrange interpolation method to interpolate the gene data; S23, searching for a multi-label Markov blanket MB, and using a multi-label causal feature selection learning method MLMB to learn the class variable MB; step S23 specifically includes: (1) Mining the causal mechanism of each gene signature; from all features and all except Y i Learn each Y from the label i MB of class labels, i.e. from V\Y i Find Y i MB, and record the result as MB(Y i ); (2) Detect features that are ignored due to label correlation; specifically, from F\MB(Y i ) detects features that are ignored due to strong label correlation; according to the features and Y i The correlation degree is sorted in descending order, and the top k1% features with the highest correlation are selected and stored in sel(Y i ) to detect, selectively from sel(Y i ) to restore the lost true MB features; (3) Correct the false features obtained; specifically, according to MB(Y i ) and Y i The correlation degree is arranged in ascending order, and the top k2% features with the weakest correlation are selected and stored in SelFea(Y i ) in; traverse q tags and detect SelFea(Y i ) whether the MB of the feature contains Y i , from SelFea(Y i ) to remove false MB features; S24, obtaining an extended-MB; the extended-MB is composed of the union of the MB of the class variable and the MB of each variable in the class variable MB; S25, updating the data set; updating the data set to a new data set containing only extended-MB; S26, determine whether the iteration termination condition is met, if so, stop, if not, repeat steps S22-S25 continuously; S27, the iteration ends and the feature subset of the gene expression data set is returned; S3, training a gene classification model using a feature subset of the gene expression data set to obtain a trained gene classification model; S4. Obtain gene data to be classified, and classify the genes to be classified using a trained gene classification model based on the feature subset of the gene data to be classified to obtain labels for the gene data to be classified.

2. The method according to claim 1, characterized in that In the gene expression dataset, F = {F1, F2, ..., F m } represents an m-dimensional feature set, Y = {Y1, Y2, ..., Y q } represents a q-dimensional label set, V = F ∪ {Y}, s is any set of variables in V; (where i≠j, ) indicates V i Conditionally independent of V given S j ; V\Y i With V\{Y i } is equivalent to, which means that in V, Y is divided i All features except 3. The method according to claim 1, characterized in that The gene data interpolation described in step S22 specifically includes: using the observation information of the incomplete instances and the complete instances in the gene expression data set to estimate the missing values, and using the KNN interpolation method to interpolate the gene data.

4. The method according to claim 1, characterized in that: The gene classification model adopts MLKNN.

5. A gene classification system based on multi-label feature selection, characterized in that: The gene classification system executes the gene classification method based on multi-label feature selection as claimed in claim 1, comprising: a multi-label feature selection model construction module, a feature selection module, a gene classification model training module, and a gene classification module; Multi-label feature selection model construction module, which constructs a multi-label feature selection model based on missing gene data; A feature selection module is used to obtain a gene expression data set, and perform feature selection on the gene expression data set using the multi-label feature selection model based on missing gene data to obtain a feature subset of the gene expression data set; A gene classification model training module is used to train a gene classification model using a feature subset of the gene expression data set to obtain a trained gene classification model; The gene classification module obtains the gene data to be classified, and based on the feature subset of the gene data to be classified, classifies the gene to be classified using a trained gene classification model to obtain a label of the gene data to be classified.

6. A computer device, characterized in that: The device comprises a memory and a processor, wherein a computer program is stored in the memory, and the processor implements the method according to any one of claims 1 to 4 when executing the computer program.

Images