Population data feature extraction method, device and terminal equipment

Abstract

The present invention provides a method, apparatus, and terminal device for extracting features from population data. The method comprises: obtaining population data from a target population monitoring system; inputting non-text data from each population data set into a preset feature extraction model to obtain a non-text feature vector corresponding to each population data set; extracting valid keywords from the text data from each population data set to obtain a valid keyword group corresponding to each population data set; performing fuzzy conversion on the valid keyword group to obtain a text feature vector corresponding to each population data set; and determining a feature extraction result corresponding to each population data set based on the non-text feature vector and the text feature vector corresponding to each population data set. The method, apparatus, and terminal device for extracting features from population data provided by the present invention can more accurately and comprehensively extract features from population data.

Description

Technical Field

[0001] The present invention belongs to the field of data processing technology, and more specifically, relates to a method, device and terminal equipment for extracting population data features. Background Art

[0002] With the rapid development of Internet+ and big data technology, various massive data are expanding rapidly, and the era of big data has quietly arrived. Its new concepts and application technologies have brought unique development prospects for government intelligent services, and population monitoring systems have also come into being.

[0003] It is known that a large amount of population data is stored in the population monitoring system, and the data type of the population data is complex. Therefore, how to more accurately extract the data features of the population data to facilitate subsequent data processing has become a problem that technical personnel in this field urgently need to solve. Summary of the Invention

[0004] The purpose of the present invention is to provide a method, apparatus and terminal device for extracting features from population data, so as to improve the accuracy of feature extraction from population data.

[0005] According to a first aspect of an embodiment of the present invention, a method for extracting features from population data is provided. The method is applied to a population monitoring system and includes:

[0006] Obtaining population data for target population monitoring systems;

[0007] Input the non-text data in each demographic data into a preset feature extraction model to obtain the non-text feature vector corresponding to each demographic data;

[0008] Extracting effective keywords from text data in each demographic data to obtain effective keyword groups corresponding to each demographic data, and performing fuzzy conversion on the effective keyword groups to obtain text feature vectors corresponding to each demographic data;

[0009] The feature extraction results corresponding to the respective population data are determined based on the non-text feature vectors corresponding to the respective population data and the text feature vectors corresponding to the respective population data.

[0010] According to a second aspect of an embodiment of the present invention, a population data feature extraction device is provided. The population data feature extraction device is applied to a population monitoring system, and the population data feature extraction device includes:

[0011] A data acquisition module, used to obtain population data of a target population monitoring system;

[0012] A first feature extraction module is used to input the non-text data in each population data into a preset feature extraction model to obtain the non-text feature vector corresponding to each population data;

[0013] A second feature extraction module is used to extract effective keywords from the text data of each population data to obtain effective keyword groups corresponding to each population data, and perform fuzzy conversion on the effective keyword groups to obtain text feature vectors corresponding to each population data;

[0014] The feature fusion module is used to determine the feature extraction results corresponding to each population data based on the non-text feature vector corresponding to each population data and the text feature vector corresponding to each population data.

[0015] According to a third aspect of an embodiment of the present invention, a terminal device is provided, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the above-mentioned method for extracting demographic data features when executing the computer program.

[0016] According to a fourth aspect of an embodiment of the present invention, a computer-readable storage medium is provided, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the above-mentioned method for extracting characteristics of population data are implemented.

[0017] The beneficial effects of the population data feature extraction method, apparatus, and terminal device provided by the embodiments of the present invention are:

[0018] Unlike existing approaches that directly extract features, this method distinguishes between textual and non-textual data within population data, extracting feature vectors for each. Finally, the feature extraction results for the population data are determined based on these vectors. This means that compared to existing approaches, this method can more comprehensively and accurately describe population data, achieving more accurate data feature extraction. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0020] Figure 1 A schematic flow chart of a method for extracting demographic data features according to an embodiment of the present invention;

[0021] Figure 2A structural block diagram of a population data feature extraction device provided by one embodiment of the present invention;

[0022] Figure 3 A schematic block diagram of a terminal device provided in one embodiment of the present invention. DETAILED DESCRIPTION

[0023] In order to make the technical problems, technical solutions and beneficial effects to be solved by the present invention more clearly understood, the present invention is further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0024] Please refer to Figure 1 , Figure 1 This is a flow chart of a method for extracting features from population data provided by one embodiment of the present invention. The method is applied to a population monitoring system and includes:

[0025] S101: Obtain population data of a target population monitoring system.

[0026] S102: Inputting the non-text data in each demographic data into a preset feature extraction model to obtain a non-text feature vector corresponding to each demographic data.

[0027] S103: extracting effective keywords from the text data of each population data to obtain effective keyword groups corresponding to each population data, performing fuzzy conversion on the effective keyword groups corresponding to each population data to obtain text feature vectors corresponding to each population data.

[0028] In this embodiment, the effective keyword groups are fuzzy converted to obtain text feature vectors corresponding to various demographic data, which can be described in detail as follows:

[0029] For a valid keyword group corresponding to a certain population data, each valid keyword in the valid keyword group is fuzzy-converted to obtain a fuzzy quantity corresponding to each valid keyword, and the fuzzy quantities corresponding to each valid keyword are combined to obtain a text feature vector corresponding to the population data.

[0030] S104: Determine a feature extraction result corresponding to each demographic data based on the non-text feature vector corresponding to each demographic data and the text feature vector corresponding to each demographic data.

[0031] In this embodiment, the feature extraction results corresponding to each demographic data are determined based on the non-text feature vectors corresponding to each demographic data and the text feature vectors corresponding to each demographic data, which can be described in detail as follows:

[0032] The non-text feature vectors corresponding to each population data and the text feature vectors corresponding to each population data are combined, and the combined feature vectors corresponding to each population data are used as feature extraction results corresponding to each population data.

[0033] As can be seen from the above, unlike existing solutions that directly perform feature extraction, the present invention distinguishes between textual and non-textual data in population data, extracts feature vectors for each, and finally determines the feature extraction results for the population data based on the textual and non-textual feature vectors. In other words, compared to existing technologies, the present invention can more comprehensively and accurately describe population data and achieve more accurate data feature extraction.

[0034] Optionally, as a specific implementation of the population data feature extraction method provided in an embodiment of the present invention, the feature extraction model training method is:

[0035] Obtain population sample data, and input the population sample data into a pre-built feature extraction network to train and obtain a first feature extraction model.

[0036] Extracting the initialization weight coefficient of the feature extraction network and the weight coefficient of the first feature extraction model, and determining the preferred weight coefficient based on the initialization weight coefficient of the feature extraction network and the weight coefficient of the first feature extraction model.

[0037] The value of the preferred weight coefficient in the first feature extraction model is randomly increased according to a preset ratio to obtain multiple second feature extraction models, and the population sample data is input into the multiple second feature extraction models to obtain multiple groups of auxiliary sample data.

[0038] The first feature extraction model is trained twice based on the population sample data and multiple groups of auxiliary sample data to obtain a trained feature extraction model.

[0039] In this embodiment, the randomly increased preferred weight coefficients are different, and the obtained second feature extraction models are also different. Therefore, the value of the preferred weight coefficient in the first feature extraction model can be randomly increased according to a preset ratio to obtain multiple second feature extraction models.

[0040] In this embodiment, the preferred weight coefficient refers to a weight coefficient that has a greater impact on feature extraction accuracy.

[0041] In this embodiment, multiple groups of auxiliary sample data are auxiliary samples generated based on the change of preferred weight coefficients. Introducing the auxiliary sample data into the training of the first feature extraction model can effectively improve the diversity of sample data, thereby improving the training accuracy of the first feature extraction model.

[0042] Optionally, as a specific implementation of the population data feature extraction method provided in an embodiment of the present invention, determining the preferred weight coefficient based on the initialization weight coefficient of the feature extraction network and the weight coefficient of the first feature extraction model includes:

[0043] The rate of change of the weight coefficient is determined based on the initialized weight coefficient of the feature extraction network and the weight coefficient of the first feature extraction model.

[0044] The weight coefficient with a change rate greater than a preset threshold is used as the preferred weight coefficient.

[0045] In this embodiment, the rate of change of the weight coefficient is the ratio of the difference (herein, the absolute value) between the weight coefficient of the first feature extraction model and the initialization weight coefficient of the feature extraction network to the initialization weight coefficient of the feature extraction network.

[0046] Optionally, as a specific implementation of the population data feature extraction method provided in an embodiment of the present invention, the first feature extraction model is trained twice based on the population sample data and multiple sets of auxiliary sample data to obtain a trained feature extraction model, including:

[0047] The population sample data and multiple groups of auxiliary sample data are combined to obtain combined sample data, and the first feature extraction model is trained twice based on the combined sample data to obtain a trained feature extraction model.

[0048] In this embodiment, the population sample data and multiple groups of auxiliary sample data can be randomly combined to obtain combined sample data, and the first feature extraction model can be trained twice based on the combined sample data to obtain a trained feature extraction model.

[0049] Optionally, as a specific implementation of the method for extracting demographic data features provided by an embodiment of the present invention, for text data in certain demographic data, a method for extracting effective keywords from the text data is as follows:

[0050] An approximate search is performed in a preset keyword library to extract candidate keywords in the text data and target keywords corresponding to each candidate keyword in the keyword library.

[0051] Each candidate keyword and the target keyword corresponding to each candidate keyword are input into a preset matching model to determine the valid keyword corresponding to the text data.

[0052] In this embodiment, the valid keywords corresponding to the text data are the valid keywords corresponding to the population data, and all valid keywords corresponding to the text data constitute the valid keyword group corresponding to the population data.

[0053] In this embodiment, the text data in a certain population data may be a text data set describing the current status of people in the target population monitoring system.

[0054] In this embodiment, a candidate keyword is a keyword that is similar to a target keyword in the keyword library. For example, if a target keyword is "inability to act", and the text data contains a phrase "lack of ability to act" or "x action x ability" (where x represents a character or word), then during an approximate search, "inability to act" will become the target keyword corresponding to "lack of ability to act" or "x action x ability" (where x represents a character or word). In other words, an approximate search is to search the keyword library for target keywords that have overlapping phrases with candidate keywords. It is worth noting that a candidate keyword is a keyword that already corresponds to a target keyword. If there is a keyword in the population data that does not match any target keyword in the keyword library, then the keyword cannot be called a candidate keyword.

[0055] Optionally, as a specific implementation of the population data feature extraction method provided in an embodiment of the present invention, the preset matching model is a probabilistic neural network model.

[0056] Inputting each candidate keyword and the target keyword corresponding to each candidate keyword into a preset matching model to determine the valid keyword corresponding to the text data includes:

[0057] S41: Select any candidate keyword, input the candidate keyword and the target keyword corresponding to the candidate keyword into a preset matching model, and obtain a matching probability between the candidate keyword and the target keyword.

[0058] S42: If the matching probability is greater than the preset probability value, the target keyword is used as a valid keyword. If the matching probability is not greater than the preset probability value, the candidate keyword is deleted.

[0059] S43: If all candidate keywords have been traversed, each valid keyword obtained in step S42 is used as a valid keyword corresponding to the text data.

[0060] If not all candidate keywords have been traversed, the process returns to step S41.

[0061] In this embodiment, a probabilistic neural network model is used to receive a candidate keyword and a target keyword corresponding to the candidate keyword and output a matching probability between the two. If the matching probability is not greater than a preset probability value, the match fails and the candidate keyword is deleted. If the matching probability is greater than the preset probability value, the match succeeds and the target keyword is used as the valid keyword representing the candidate keyword, facilitating subsequent calculations.

[0062] Corresponding to the population data feature extraction method in the above embodiment, Figure 2 This is a structural block diagram of a population data feature extraction device provided by an embodiment of the present invention. For ease of explanation, only the parts related to the embodiment of the present invention are shown. Figure 2 The population data feature extraction device 20 is applied to a population monitoring system. The population data feature extraction device 20 includes: a data acquisition module 21, a first feature extraction module 22, a second feature extraction module 23, and a feature fusion module 24.

[0063] The data acquisition module 21 is used to acquire population data of the target population monitoring system.

[0064] The first feature extraction module 22 is used to input the non-text data in each demographic data into a preset feature extraction model to obtain the non-text feature vector corresponding to each demographic data.

[0065] The second feature extraction module 23 is used to extract effective keywords from the text data of each demographic data, obtain effective keyword groups corresponding to each demographic data, perform fuzzy conversion on the effective keyword groups, and obtain text feature vectors corresponding to each demographic data.

[0066] The feature fusion module 24 is configured to determine a feature extraction result corresponding to each demographic data based on the non-text feature vector corresponding to each demographic data and the text feature vector corresponding to each demographic data.

[0067] Alternatively, refer to Figure 2 As a specific implementation of the population data feature extraction device provided in an embodiment of the present invention, the population data feature extraction device 20 may further include a model training module 25, which is used to:

[0068] Obtain population sample data, and input the population sample data into a pre-built feature extraction network to train and obtain a first feature extraction model.

[0069] Extracting the initialization weight coefficient of the feature extraction network and the weight coefficient of the first feature extraction model, and determining the preferred weight coefficient based on the initialization weight coefficient of the feature extraction network and the weight coefficient of the first feature extraction model.

[0070] The value of the preferred weight coefficient in the first feature extraction model is randomly increased according to a preset ratio to obtain multiple second feature extraction models, and the population sample data is input into the multiple second feature extraction models to obtain multiple groups of auxiliary sample data.

[0071] The first feature extraction model is trained twice based on the population sample data and multiple groups of auxiliary sample data to obtain a trained feature extraction model.

[0072] Optionally, as a specific implementation of the population data feature extraction device provided in an embodiment of the present invention, determining the preferred weight coefficient based on the initialization weight coefficient of the feature extraction network and the weight coefficient of the first feature extraction model includes:

[0073] The rate of change of the weight coefficient is determined based on the initialized weight coefficient of the feature extraction network and the weight coefficient of the first feature extraction model.

[0074] The weight coefficient with a change rate greater than a preset threshold is used as the preferred weight coefficient.

[0075] Optionally, as a specific implementation of the population data feature extraction device provided in an embodiment of the present invention, the first feature extraction model is trained twice based on the population sample data and multiple sets of auxiliary sample data to obtain a trained feature extraction model, including:

[0076] The population sample data and multiple groups of auxiliary sample data are combined to obtain combined sample data, and the first feature extraction model is trained twice based on the combined sample data to obtain a trained feature extraction model.

[0077] Optionally, as a specific implementation of the demographic data feature extraction device provided in an embodiment of the present invention, for text data in certain demographic data, a method for extracting effective keywords from the text data is as follows:

[0078] An approximate search is performed in a preset keyword library to extract candidate keywords in the text data and target keywords corresponding to each candidate keyword in the keyword library.

[0079] Each candidate keyword and the target keyword corresponding to each candidate keyword are input into a preset matching model to determine the valid keyword corresponding to the text data.

[0080] Optionally, as a specific implementation of the population data feature extraction device provided in an embodiment of the present invention, determining a feature extraction result corresponding to each population data based on a non-text feature vector corresponding to each population data and a text feature vector corresponding to each population data includes:

[0081] The non-text feature vectors corresponding to each population data and the text feature vectors corresponding to each population data are combined, and the combined feature vectors corresponding to each population data are used as feature extraction results corresponding to each population data.

[0082] See also Figure 3 , Figure 3 This is a schematic block diagram of a terminal device provided by an embodiment of the present invention. Figure 3The terminal 300 in the embodiment shown may include: one or more processors 301, one or more input devices 302, one or more output devices 303, and one or more memories 304. The processors 301, input devices 302, output devices 303, and memories 304 communicate with each other via a communication bus 305. The memory 304 is used to store computer programs, which include program instructions. The processor 301 is used to execute the program instructions stored in the memory 304. The processor 301 is configured to call the program instructions to perform the following operations to perform the functions of the modules / units in the above-mentioned device embodiments, such as Figure 2 The functions of modules 21 to 25 are shown.

[0083] It should be understood that in the embodiment of the present invention, the processor 301 may be a central processing unit (CPU), and 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. A general-purpose processor may be a microprocessor or any conventional processor, etc.

[0084] The input device 302 may include a touchpad, a fingerprint sensor (for collecting the user's fingerprint information and fingerprint direction information), a microphone, etc. The output device 303 may include a display (LCD, etc.), a speaker, etc.

[0085] The memory 304 may include a read-only memory and a random access memory, and provides instructions and data to the processor 301. A portion of the memory 304 may also include a non-volatile random access memory. For example, the memory 304 may also store information about the device type.

[0086] In a specific implementation, the processor 301, input device 302, and output device 303 described in the embodiment of the present invention can execute the implementation methods described in the first and second embodiments of the population data feature extraction method provided in the embodiment of the present invention, and can also execute the implementation methods of the terminal described in the embodiment of the present invention, which will not be repeated here.

[0087] In another embodiment of the present invention, a computer-readable storage medium is provided. The computer-readable storage medium stores a computer program. The computer program includes program instructions. When the program instructions are executed by a processor, all or part of the process in the above-mentioned embodiment method is implemented. The computer program can also be used to instruct related hardware to complete the process. The computer program can be stored in a computer-readable storage medium. When the computer program is executed by the processor, the steps of each of the above-mentioned method embodiments are implemented. The computer program includes computer program code. The computer program code can be in source code form, object code form, executable file or some intermediate form. The computer-readable medium can include: any entity or device capable of carrying computer program code, recording medium, USB flash drive, mobile hard disk, magnetic disk, optical disk, computer memory, read-only memory (ROM), random access memory (RAM), electric carrier signal, telecommunication signal and software distribution medium. It should be noted that the content contained in the computer-readable medium can be appropriately increased or decreased according to the requirements of legislation and patent practice in the jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, computer-readable media do not include electric carrier signals and telecommunication signals.

[0088] The computer-readable storage medium may be an internal storage unit of the terminal in any of the aforementioned embodiments, such as a hard disk or memory of the terminal. The computer-readable storage medium may also be an external storage device of the terminal, such as a plug-in hard disk equipped on the terminal, a smart memory card (SMC), a secure digital (SD) card, a flash memory card, etc. Furthermore, the computer-readable storage medium may include both an internal storage unit of the terminal and an external storage device. The computer-readable storage medium is used to store computer programs and other programs and data required by the terminal. The computer-readable storage medium may also be used to temporarily store data that has been output or is about to be output.

[0089] Those skilled 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 above description has generally described the composition and steps of each example according to function. 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.

[0090] Those skilled in the art will clearly understand that, for the convenience and brevity of description, the specific working processes of the terminals and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0091] In the several embodiments provided in this application, it should be understood that the disclosed terminals and methods can be implemented in other ways. For example, the device embodiments described above are merely illustrative. For example, the division of units is merely a logical function division. In actual implementation, there may be other division methods, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. In addition, the mutual coupling or direct coupling or communication connection shown or discussed can be an indirect coupling or communication connection through some interfaces or units, or can be an electrical, mechanical or other form of connection.

[0092] Units described as separate components may or may not be physically separate, and components shown as units may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of these units may be selected based on actual needs to achieve the objectives of the embodiments of the present invention.

[0093] In addition, the functional units in the various embodiments of the present invention may be integrated into a single processing unit, each unit may exist physically separately, or two or more units may be integrated into a single unit. The aforementioned integrated units may be implemented in the form of hardware or software functional units.

[0094] The above are merely specific embodiments of the present invention, but the scope of protection of the present invention is not limited thereto. Any person skilled in the art can easily conceive of various equivalent modifications or substitutions within the technical scope disclosed in the present invention, and such modifications or substitutions are intended to be within the scope of protection of the present invention. Therefore, the scope of protection of the present invention shall be subject to the scope of protection of the claims.

Claims

1. A method for extracting features from population data, characterized in that: The population data feature extraction method is applied to a population monitoring system, and the method comprises: Obtaining population data for target population monitoring systems; Input the non-text data in each demographic data into a preset feature extraction model to obtain the non-text feature vector corresponding to each demographic data; For any text data in the population data, an approximate search is performed in a preset keyword library to extract candidate keywords in the text data and target keywords corresponding to each candidate keyword in the keyword library; S41: Select a candidate keyword, input the candidate keyword and the target keyword corresponding to the candidate keyword into a preset matching model, and obtain a matching probability between the candidate keyword and the target keyword; S42: If the matching probability is greater than the preset probability value, the target keyword is used as a valid keyword; if the matching probability is not greater than the preset probability value, the candidate keyword is deleted; S43: If all candidate keywords have been traversed, then each valid keyword obtained in step S42 is used as a valid keyword corresponding to the text data; if not all candidate keywords have been traversed, then return to step S41 to obtain a valid keyword group corresponding to each population data, and for any valid keyword group, perform fuzzy conversion on each valid keyword in the valid keyword group to obtain a fuzzy quantity corresponding to each valid keyword, and combine the fuzzy quantities corresponding to each valid keyword to obtain a text feature vector corresponding to the population data; wherein, the valid keyword group corresponding to each population data is a set of all valid keywords in the text data of each population data; The feature extraction results corresponding to the respective population data are determined based on the non-text feature vectors corresponding to the respective population data and the text feature vectors corresponding to the respective population data.

2. The method for extracting demographic data features according to claim 1, wherein: The training method of the feature extraction model is: Obtaining population sample data, and inputting the population sample data into a pre-built feature extraction network to train and obtain a first feature extraction model; Extracting an initialization weight coefficient of the feature extraction network and a weight coefficient of the first feature extraction model, and determining a preferred weight coefficient based on the initialization weight coefficient of the feature extraction network and the weight coefficient of the first feature extraction model; Randomly increasing the value of the preferred weight coefficient in the first feature extraction model according to a preset ratio to obtain multiple second feature extraction models, and inputting the population sample data into the multiple second feature extraction models to obtain multiple sets of auxiliary sample data; The first feature extraction model is trained twice based on the population sample data and the multiple groups of auxiliary sample data to obtain a trained feature extraction model.

3. The method for extracting demographic data features according to claim 2, wherein: The determining of the preferred weight coefficient based on the initialization weight coefficient of the feature extraction network and the weight coefficient of the first feature extraction model includes: Determining a rate of change of a weight coefficient based on an initialized weight coefficient of the feature extraction network and a weight coefficient of the first feature extraction model; The weight coefficient with a change rate greater than a preset threshold is used as the preferred weight coefficient.

4. The method for extracting demographic data features according to claim 2, wherein: The second training of the first feature extraction model based on the population sample data and the multiple sets of auxiliary sample data to obtain a trained feature extraction model includes: The population sample data and the multiple groups of auxiliary sample data are combined to obtain combined sample data, and the first feature extraction model is trained twice based on the combined sample data to obtain a trained feature extraction model.

5. The method for extracting demographic data features according to claim 1, wherein: For text data in a certain population data, the method for extracting effective keywords from the text data is: Performing an approximate search in a preset keyword library to extract candidate keywords in the text data and target keywords corresponding to each candidate keyword in the keyword library; Each candidate keyword and the target keyword corresponding to each candidate keyword are input into a preset matching model to determine the valid keyword corresponding to the text data.

6. The method for extracting demographic data features according to claim 1, wherein: The determining of the feature extraction results corresponding to each population data based on the non-text feature vector corresponding to each population data and the text feature vector corresponding to each population data includes: The non-text feature vectors corresponding to each population data and the text feature vectors corresponding to each population data are combined, and the combined feature vectors corresponding to each population data are used as feature extraction results corresponding to each population data.

7. A population data feature extraction device, characterized in that: The population data feature extraction device is applied to a population monitoring system, and the population data feature extraction device includes: A data acquisition module, used to obtain population data of a target population monitoring system; A first feature extraction module is used to input the non-text data in each population data into a preset feature extraction model to obtain the non-text feature vector corresponding to each population data; A second feature extraction module is used to perform an approximate search in a preset keyword library for text data in any population data, and extract candidate keywords in the text data and target keywords corresponding to each candidate keyword in the keyword library; S41: Select a candidate keyword, input the candidate keyword and the target keyword corresponding to the candidate keyword into a preset matching model, and obtain a matching probability between the candidate keyword and the target keyword; S42: If the matching probability is greater than the preset probability value, the target keyword is used as a valid keyword; if the matching probability is not greater than the preset probability value, the candidate keyword is deleted; S43: If all candidate keywords have been traversed, then each valid keyword obtained in step S42 is used as a valid keyword corresponding to the text data; if not all candidate keywords have been traversed, then return to step S41 to obtain a valid keyword group corresponding to each population data, and for any valid keyword group, perform fuzzy conversion on each valid keyword in the valid keyword group to obtain a fuzzy quantity corresponding to each valid keyword, and combine the fuzzy quantities corresponding to each valid keyword to obtain a text feature vector corresponding to the population data; wherein, the valid keyword group corresponding to each population data is a set of all valid keywords in the text data of each population data; The feature fusion module is used to determine the feature extraction results corresponding to each population data based on the non-text feature vector corresponding to each population data and the text feature vector corresponding to each population data.

8. The population data feature extraction device according to claim 7, wherein: The invention also includes a model training module, wherein the model training module is used to obtain population sample data and input the population sample data into a pre-built feature extraction network to train and obtain a first feature extraction model; Extracting an initialization weight coefficient of the feature extraction network and a weight coefficient of the first feature extraction model, and determining a preferred weight coefficient based on the initialization weight coefficient of the feature extraction network and the weight coefficient of the first feature extraction model; Randomly increasing the value of the preferred weight coefficient in the first feature extraction model according to a preset ratio to obtain multiple second feature extraction models, and inputting the population sample data into the multiple second feature extraction models to obtain multiple sets of auxiliary sample data; The first feature extraction model is trained twice based on the population sample data and the multiple groups of auxiliary sample data to obtain a trained feature extraction model.

9. A terminal device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 6 are implemented.

10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.

Images