Service matching method, system and equipment based on feature extraction and medium
Through feature extraction and classification neural network technology, the existing service matching methods are solved in terms of accuracy, and the efficient matching of user needs and service applicability is achieved, and the user experience is improved.
Patent Information
- Application Number
- CN202411869757.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-12-18
- Publication Date
- 2025-04-22
AI Technical Summary
The existing service matching method is not ideal in terms of matching accuracy, and cannot effectively match user needs and meet the applicable conditions of the service, resulting in users being unable to truly enjoy the matching services.
Through feature extraction technology, natural language text is converted into abstract layer vectorized features, and classification neural networks are used to classify service sets. According to the service set classification, conditional question-and-answer options are provided, allowing users to independently choose to match service information that meets their own needs.
It improves the accuracy of service matching, enables users to find service information that truly meets their own situation and enjoyment conditions, and ensures the applicability of the service and user satisfaction.
Smart Images
Figure CN119862256A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of feature engineering and service matching, and in particular to a service matching method, system, device and medium based on feature extraction. Background Art
[0002] With the continuous improvement of service awareness in all walks of life, the types and quantity of services are increasing, and the fields involved are becoming wider and wider. Excellent service is an important way to improve the user's happiness index. In order to enable users to better understand various services and to enable services to more effectively serve those in need, the reasonable matching and promotion of services has become an important and key means of promotion.
[0003] Computers cannot directly process unstructured text data, and it is also difficult to directly interpret and understand natural language, and the final effect is not good. At present, the service matching method has not been particularly ideal in terms of matching accuracy. On the one hand, it may not match the service information required by the user. More importantly, on the other hand, the conditions for enjoying the service are generally strict and precise, and the user may not really enjoy the matched service information.
[0004] Therefore, how to provide users with service information that matches their needs and meets applicable conditions to help users better enjoy service convenience is a technical problem that needs to be solved urgently. Summary of the invention
[0005] The technical task of the present invention is to provide a service matching method, system, device and medium based on feature extraction to solve the problem of how to provide users with service information that matches their needs and meets applicable conditions, so as to help users better enjoy service convenience.
[0006] The technical task of the present invention is achieved in the following way: a service matching method based on feature extraction, the method is as follows:
[0007] Collect user basic information and demand text as initial text information, and build an initial service set based on the initial text information;
[0008] Extract features from text information, vectorize features of the natural language text abstraction layer, and classify the service set;
[0009] According to the classification of service sets, relevant condition options are preset for users to choose from, that is, conditional questions and answers are provided, and services are further matched based on the user's selection results, so that users can independently match relevant service information that suits them and ensure the applicability of the service.
[0010] Preferably, the initial text information is obtained by a method of text input, document import, or voice input;
[0011] Among them, the text input method directly inputs relevant text information on the page; the document import method can import documents in word and xlsx formats and convert them into corresponding text information; the voice input method obtains audio through the sound receiving device, and then obtains the corresponding text information after uploading audio files, data preprocessing, data enhancement of the original audio, Mel-frequency cepstral coefficients, spectrum graph data enhancement and deep neural network processing and decoding.
[0012] Preferably, the initial text information is a natural language description including the user's basic information and user demand information; wherein, the user's basic information is used to describe the user's gender, age, education, work experience and physical condition; the user's demand information is used to describe the user's current needs for scene description and event description.
[0013] Preferably, feature extraction is performed on the text information, the features of the natural language text abstraction layer are vectorized, and the service set classification is performed as follows:
[0014] Data preprocessing: Perform preprocessing operations such as data loading, data cleaning, word segmentation, and corpus establishment on the initial text information to obtain the preprocessed text information;
[0015] Build the model: Use one-hot encoding, TD-IDF text feature extraction model or word2vector word vector extraction model to extract text features, use AlexNet, VGGNet or ResNet classification neural network algorithm to classify services, initialize the model, define training and evaluation functions, and train the model to obtain the trained model;
[0016] Text matching service classification: obtain the corresponding service classification through the trained model.
[0017] Preferably, the model is constructed as follows:
[0018] Convert the preprocessed segmented text information into one-hot encoding wi;
[0019] The word segmentation code wi is converted into the corresponding hidden layer vector through the weight matrix W;
[0020] Convert the hidden layer vector to the word vector output layer through another weight matrix W';
[0021] The word vector is passed into the neural network model as the input layer for classification;
[0022] The classification result is obtained after normalization by the softmax function; among them, the loss function optimized by the neural network model uses cross entropy.
[0023] Preferably, before providing conditional questions and answers, the service information is structured, and the service text is structured into pieces of judgeable conditional information and stored in a database; when querying, accurate matching is performed in the database based on the user's answer.
[0024] A service matching system based on feature extraction, the system comprising:
[0025] A collection unit, used to collect basic information and demand text of the user as initial text information, and to construct an initial service set through the initial text information;
[0026] The feature extraction and classification unit is used to extract features from text information, vectorize the features of the natural language text abstraction layer, and classify the service set;
[0027] The matching unit is used to preset relevant condition options for users to choose according to the service set classification, that is, to provide conditional questions and answers, and further match services according to the user's selection results, so that users can independently match relevant service information that suits them and ensure the applicability of the service.
[0028] Preferably, the collecting unit comprises:
[0029] The data preprocessing module is used to perform preprocessing operations such as data loading, data cleaning, word segmentation, and corpus establishment on the initial text information to obtain the preprocessed text information;
[0030] Model building module, used to extract text features using one-hot encoding, TD-IDF text feature extraction model or word2vector word vector extraction model, use AlexNet, VGGNet or ResNet classification neural network algorithm to classify services, initialize the model, define training and evaluation functions and model training, and obtain the trained model;
[0031] The text matching service classification module is used to obtain the corresponding service classification through the trained model;
[0032] The working process of the model building module is as follows:
[0033] ①Convert the preprocessed and segmented text information into one-hot encoding wi;
[0034] ②Convert the segmented code wi into the corresponding hidden layer vector through the weight matrix W;
[0035] ③ Convert the hidden layer vector into the word vector output layer through another weight matrix W';
[0036] ④ Pass the word vector as the input layer into the neural network model for classification;
[0037] ⑤ The classification result is obtained after normalization by the softmax function; among them, the loss function optimized by the neural network model uses cross entropy.
[0038] An electronic device comprising: a memory and at least one processor;
[0039] Wherein, the memory stores computer-executable instructions;
[0040] The at least one processor executes the computer-executable instructions stored in the memory, so that the at least one processor performs the service matching method based on feature extraction as described above.
[0041] A computer-readable storage medium stores computer-executable instructions. When a processor executes the computer-executable instructions, the service matching method based on feature extraction as described above is implemented.
[0042] The service matching method, system, device and medium based on feature extraction of the present invention have the following advantages:
[0043] (1) The present invention supports extracting feature vectors of personal information and demand texts input by users through feature extraction algorithms such as word vectors, supports screening matching service sets through algorithms such as classification neural networks, supports extracting conditional questions based on services in service sets, and supports accurately matching matching services based on question and answer results provided by users, so that users can independently match relevant service information that matches themselves, thereby improving the awareness of services and enabling the public to better enjoy the benefits brought by services;
[0044] (ii) The feature extraction of the present invention processes and converts noisy, unstructured text data into a structured, vectorized format that can be understood by any machine learning algorithm, thereby achieving the purpose of classification, matching, etc.;
[0045] (3) The present invention helps users find service information that is relevant to their own situation and that actually meets the conditions for enjoying the service, and helps the service to truly play its role and serve those who need the service;
[0046] (iv) The text input of the present invention supports multiple input modes, including typing input, document import, voice input, etc. Users can choose a suitable text input mode, which reduces the difficulty of use and is convenient to use;
[0047] (V) The present invention uses a feature extraction algorithm and a classification neural network to parse the input text and classify the matching service set, determine the service matching direction, and reduce unnecessary service matching;
[0048] (VI) The present invention generates corresponding conditional questions and answers for users to fill in based on the classified service sets, further confirms the applicability of the service, reduces the occurrence of service mismatching, and improves the quality of the service;
[0049] (VII) The present invention supports users to describe their initial situation and answer the corresponding conditional questions of the system, and provides users with service information that matches their needs and meets applicable conditions, helping users to better enjoy the convenience of services, and supports multiple text input methods to facilitate users to describe their situations;
[0050] (VIII) The present invention utilizes feature extraction methods and classification neural networks to narrow the scope of the service set, reduce the number of questions and answers that users need to fill in, and facilitate user use;
[0051] (IX) The present invention provides corresponding precise condition matching options based on structured service information to ensure the applicability of the service. BRIEF DESCRIPTION OF THE DRAWINGS
[0052] The present invention is further described below in conjunction with the accompanying drawings.
[0053] Attached Figure 1 The flowchart of the service matching method based on feature extraction is shown in FIG. DETAILED DESCRIPTION
[0054] The service matching method, system, device and medium based on feature extraction of the present invention are described in detail below with reference to the accompanying drawings and specific embodiments of the specification.
[0055] Embodiment 1:
[0056] As attached Figure 1 As shown, this embodiment provides a service matching method based on feature extraction, and the method is specifically as follows:
[0057] S1. Collect user basic information and demand text as initial text information, and build an initial service set based on the initial text information;
[0058] S2, extracting features from text information, vectorizing features of the natural language text abstraction layer, and classifying the service set;
[0059] S3. According to the service set classification, relevant condition options are preset for users to choose, that is, conditional questions and answers are provided, and services are further matched according to the user's selection results, so that users can independently match relevant service information that suits them and ensure the applicability of the service.
[0060] The initial text information in step S1 of this embodiment is obtained by text input, document import, and voice input.
[0061] Among them, the text input method directly inputs relevant text information on the page; the document import method can import documents in word and xlsx formats and convert them into corresponding text information; the voice input method obtains audio through the sound receiving device, and then obtains the corresponding text information after uploading audio files, data preprocessing, data enhancement of the original audio, Mel-frequency cepstral coefficients, spectrum graph data enhancement and deep neural network processing and decoding.
[0062] The initial text information in step S1 of this embodiment is a natural language description containing the user's basic information and user demand information; wherein the user's basic information is used to describe the user's own situation such as gender, age, education, work experience and physical condition; the user's demand information is used to describe the user's current needs for scene description and event description.
[0063] In step S2 of this embodiment, the features of the text information are extracted, the features of the natural language text abstraction layer are vectorized, and the service set is classified as follows:
[0064] S201, data preprocessing: performing preprocessing operations such as data loading, data cleaning, word segmentation, and corpus establishment on the initial text information to obtain preprocessed text information;
[0065] S202, build a model: use one-hot encoding, TD-IDF text feature extraction model or word2vector word vector extraction model to extract text features, use AlexNet, VGGNet or ResNet classification neural network algorithm to classify services, initialize the model, define training and evaluation functions and model training, and obtain a trained model;
[0066] S203, text matching service classification: obtain the corresponding service classification through the trained model.
[0067] The model building in step S202 of this embodiment is specifically as follows:
[0068] S20201, converting the pre-processed segmented text information into a one-hot code wi;
[0069] S20202. Convert the segmented code wi into the corresponding hidden layer vector through the weight matrix W;
[0070] S20203, converting the hidden layer vector into a word vector output layer through another weight matrix W';
[0071] S20204, passing the word vector as the input layer into the neural network model for classification;
[0072] S20205. The classification result is obtained after normalization by the softmax function; wherein, the loss function for optimizing the neural network model uses cross entropy.
[0073] Before providing conditional questions and answers in step S3 of this embodiment, the service information is structured, and the service text is structured into pieces of judgable conditional information and stored in a database; when querying, accurate matching is performed in the database based on the user's answer.
[0074] Embodiment 2:
[0075] This embodiment provides a service matching system based on feature extraction, the system comprising:
[0076] A collection unit, used to collect basic information and demand text of the user as initial text information, and to construct an initial service set through the initial text information;
[0077] The feature extraction and classification unit is used to extract features from text information, vectorize the features of the natural language text abstraction layer, and classify the service set;
[0078] The matching unit is used to preset relevant condition options for users to choose according to the service set classification, that is, to provide conditional questions and answers, and further match services according to the user's selection results, so that users can independently match relevant service information that suits them and ensure the applicability of the service.
[0079] The collection unit in this embodiment includes:
[0080] The data preprocessing module is used to perform preprocessing operations such as data loading, data cleaning, word segmentation, and corpus establishment on the initial text information to obtain the preprocessed text information;
[0081] Model building module, used to extract text features using one-hot encoding, TD-IDF text feature extraction model or word2vector word vector extraction model, use AlexNet, VGGNet or ResNet classification neural network algorithm to classify services, initialize the model, define training and evaluation functions and model training, and obtain the trained model;
[0082] The text matching service classification module is used to obtain the corresponding service classification through the trained model.
[0083] The working process of the model building module in this embodiment is as follows:
[0084] ①Convert the preprocessed and segmented text information into one-hot encoding wi;
[0085] ②Convert the segmented code wi into the corresponding hidden layer vector through the weight matrix W;
[0086] ③ Convert the hidden layer vector into the word vector output layer through another weight matrix W';
[0087] ④ Pass the word vector as the input layer into the neural network model for classification;
[0088] ⑤ The classification result is obtained after normalization by the softmax function; among them, the loss function optimized by the neural network model uses cross entropy.
[0089] Embodiment 3:
[0090] This embodiment also provides an electronic device, including: a memory and at least one processor;
[0091] Wherein, the memory stores computer-executable instructions;
[0092] The at least one processor executes the computer-executable instructions stored in the memory, so that the at least one processor executes the service matching method based on feature extraction according to any one of the present inventions.
[0093] The processor may be a central processing unit (CPU), or 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. The processor may be a microprocessor or any conventional processor, etc.
[0094] The memory can be used to store computer programs and / or modules. The processor realizes various functions of the electronic device by running or executing the computer programs and / or modules stored in the memory, and calling the data stored in the memory. The memory can mainly include a program storage area and a data storage area, wherein the program storage area can store an operating system, at least one application required for a function, etc.; the data storage area can store data created according to the use of the terminal, etc. In addition, the memory can also include a high-speed random access memory, and can also include a non-volatile memory, such as a hard disk, a memory, a plug-in hard disk, a smart memory card (SMC), a secure digital (SD) card, a flash memory card, at least one disk storage period, a flash memory device, or other volatile solid-state storage devices.
[0095] Embodiment 4:
[0096] This embodiment also provides a computer-readable storage medium, in which a plurality of instructions are stored, and the instructions are loaded by a processor, so that the processor executes the service matching method based on feature extraction in any embodiment of the present invention. Specifically, a system or device equipped with a storage medium can be provided, on which a software program code that implements the functions of any of the above embodiments is stored, and a computer (or CPU or MPU) of the system or device reads and executes the program code stored in the storage medium.
[0097] In this case, the program code itself read from the storage medium can realize the function of any one of the above-mentioned embodiments, and thus the program code and the storage medium storing the program code constitute a part of the present invention.
[0098] The storage medium embodiments for providing the program code include a floppy disk, a hard disk, a magneto-optical disk, an optical disk (such as CD-ROM, CD-R, CD-RW, DVD-ROM, DVD-RAM, DVD-RW, DVD+RW), a magnetic tape, a non-volatile memory card, and a ROM. Alternatively, the program code can be downloaded from a server computer by a communication network.
[0099] In addition, it should be clear that the functions of any of the above embodiments can be implemented not only by executing the program code read by the computer, but also by enabling an operating system operating on the computer to complete part or all of the actual operations based on instructions from the program code.
[0100] In addition, it can be understood that the program code read from the storage medium is written to a memory provided in an expansion board inserted into the computer or written to a memory provided in an expansion unit connected to the computer, and then based on the instructions of the program code, a CPU installed on the expansion board or the expansion unit is enabled to perform part or all of the actual operations, thereby realizing the functions of any of the above-mentioned embodiments.
[0101] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or replace some or all of the technical features therein with equivalents. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.
Claims
1. A service matching method based on feature extraction, characterized in that: The method is as follows: Collect user basic information and demand text as initial text information, and build an initial service set based on the initial text information; Extract features from text information, vectorize features of the natural language text abstraction layer, and classify the service set; According to the classification of service sets, relevant condition options are preset for users to choose from, that is, conditional questions and answers are provided, and services are further matched based on the user's selection results, so that users can independently match relevant service information that suits them and ensure the applicability of the service.
2. The service matching method based on feature extraction according to claim 1, characterized in that: The initial text information is obtained by means of text input, document import, and voice input; Among them, the text input method directly inputs relevant text information on the page; the document import method can import documents in word and xlsx formats and convert them into corresponding text information; the voice input method obtains audio through the sound receiving device, and then obtains the corresponding text information after uploading audio files, data preprocessing, data enhancement of the original audio, Mel-frequency cepstral coefficients, spectrum graph data enhancement and deep neural network processing and decoding.
3. The service matching method based on feature extraction according to claim 1 or 2, characterized in that: The initial text information is a natural language description that includes the user's basic information and user demand information; the user's basic information is used to describe the user's gender, age, education, work experience and physical condition; the user's demand information is used to describe the user's current needs for scene description and event description.
4. The service matching method based on feature extraction according to claim 3 is characterized in that: Feature extraction is performed on text information, the features of the natural language text abstraction layer are vectorized, and the service set classification is performed as follows: Data preprocessing: Perform preprocessing operations such as data loading, data cleaning, word segmentation, and corpus establishment on the initial text information to obtain the preprocessed text information; Build the model: Use one-hot encoding, TD-IDF text feature extraction model or word2vector word vector extraction model to extract text features, use AlexNet, VGGNet or ResNet classification neural network algorithm to classify services, initialize the model, define training and evaluation functions, and train the model to obtain the trained model; Text matching service classification: obtain the corresponding service classification through the trained model.
5. The service matching method based on feature extraction according to claim 4 is characterized in that: The model is constructed as follows: Convert the preprocessed segmented text information into one-hot encoding w i ; The segmented code w is converted into i Convert to the corresponding hidden layer vector; Convert the hidden layer vector to the word vector output layer through another weight matrix W'; The word vector is passed into the neural network model as the input layer for classification; The classification result is obtained after normalization by the softmax function; among them, the loss function optimized by the neural network model uses cross entropy.
6. The service matching method based on feature extraction according to claim 5, characterized in that: Before providing conditional questions and answers, the service information is structured, and the service text is structured into pieces of judgable conditional information and stored in the database; when querying, accurate matching is performed in the database based on the user's answer.
7. A service matching system based on feature extraction, characterized in that: The system includes: A collection unit, used to collect basic information and demand text of the user as initial text information, and to construct an initial service set through the initial text information; The feature extraction and classification unit is used to extract features from text information, vectorize the features of the natural language text abstraction layer, and classify the service set; The matching unit is used to preset relevant condition options for users to choose according to the service set classification, that is, to provide conditional questions and answers, and further match services according to the user's selection results, so that users can independently match relevant service information that suits them and ensure the applicability of the service.
8. The service matching system based on feature extraction according to claim 7, characterized in that: The collection unit includes: The data preprocessing module is used to perform preprocessing operations such as data loading, data cleaning, word segmentation, and corpus establishment on the initial text information to obtain the preprocessed text information; Model building module, used to extract text features using one-hot encoding, TD-IDF text feature extraction model or word2vector word vector extraction model, use AlexNet, VGGNet or ResNet classification neural network algorithm to classify services, initialize the model, define training and evaluation functions and model training, and obtain the trained model; The text matching service classification module is used to obtain the corresponding service classification through the trained model; The working process of the model building module is as follows: ①Convert the preprocessed segmented text information into one-hot encoding w i ; ②The segmented code w is converted into i Convert to the corresponding hidden layer vector; ③ Convert the hidden layer vector into the word vector output layer through another weight matrix W'; ④ Pass the word vector as the input layer into the neural network model for classification; ⑤ The classification result is obtained after normalization by the softmax function; among them, the loss function optimized by the neural network model uses cross entropy.
9. An electronic device, characterized in that: include: memory and at least one processor; Wherein, the memory stores computer-executable instructions; The at least one processor executes the computer-executable instructions stored in the memory, so that the at least one processor executes the service matching method based on feature extraction as described in any one of claims 1 to 6.
10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer-executable instructions, and when the processor executes the computer-executable instructions, the service matching method based on feature extraction as described in any one of claims 1 to 6 is implemented.