210 results about "Information evaluation" patented technology

The invention belongs to the technical field of information, and relates to a multi-level dynamic intelligent automatic data processing and information evaluation method and system. The method comprises the steps that a database environment operation data set is constructed, targeted training and fine adjustment are conducted on a large language model in combination with reinforcement learning, and full-automatic data analysis, query and information extraction of multiple tables in a database are achieved; a multi-level agent interaction and cooperation framework is established, a scheduling agent is responsible for global decision making and strategy planning, an operation agent is responsible for execution of specific tasks, and the scheduling agent and the operation agent share information and coordinate the tasks to form a highly cooperative whole; and establishing a cyclic supervision and error correction mechanism, scheduling the intelligent agent to learn a normative workflow example of the target working environment and track the state of the operation intelligent agent in real time, identifying the behavior of the operation intelligent agent deviating from a predetermined path, and taking corrective measures. According to the method, high efficiency, intelligence and automation of data processing are realized, and the method has remarkable practical value and innovative significance.

The invention discloses a method for predicting abnormal indexes after thoracic surgery, and relates to the technical field of medical treatment, and the main scheme is as follows: obtaining historical data related to thoracic surgery and preoperative examination, basic information, intraoperative and postoperative recovery data of a patient, and respectively calculating a preoperative comprehensive value, a basic information evaluation value, an intraoperative risk index and a postoperative recovery index; further obtaining a predicted postoperative infection index and a predicted cardiopulmonary function index; calculating a postoperative infection index threshold value and a cardiopulmonary function index threshold value according to the historical data; by comparing a prediction index with a threshold value, triggering corresponding early warning and formulating postoperative management and intervention strategies of different levels, the postoperative rehabilitation quality and safety of a thoracic surgery patient are improved, a comprehensive method is provided for postoperative management of the thoracic surgery patient, risks can be found in time, targeted measures can be taken, and the postoperative rehabilitation safety of the thoracic surgery patient is improved. The operation success rate and the patient life quality are improved.

A data collection unit (11) collects information related to an object and sets the information as collected data (101). A data normalization unit (12) adds, as an input type, the type of data to the collected data (101) collected by the data collection unit (11), and thereby generates training data (102). A training unit (13) trains a machine learning model (103), which is a large-scale language model, on the basis of the training data (102) generated by the data normalization unit (12). By using the trained machine learning model (103) and a portion of visualization data that includes configuration information, state information, and evaluation information of the object and information indicating relationships of the object with other related objects, an inference unit (14) infers the information in the visualization data other than said portion. A correction unit (15) corrects the information in the visualization data inferred by the inference unit (14), on the basis of the collected data (101).

The invention relates to a bid invitation information feedback comparison system based on big data, and relates to the technical field of big data, and the system comprises an information acquisition module, an information processing module, an information extraction module, an information comparison module, an information evaluation module, a ranking calculation module and a decision generation module. The technical key points are as follows: bid invitation and bidding document information is collected, and a document database is constructed; and extracting qualification evaluation feature words by using a natural language processing technology, and comparing the qualification evaluation feature words with preset qualification demand information to generate a qualification bid evaluation report. In addition, the method further comprises the steps of evaluating the price rationality of the bidding party quotation information, and generating a price evaluation report according to the price rationality. And finally, calculating the ranking of the bidding parties according to each bid evaluation report, generating decision support information, and marking the bidding party with the top ranking as a recommended bidding party. Through comprehensive qualification and price evaluation, a simple, efficient and practical solution is provided for small and medium-sized enterprises.

The invention relates to the technical field of computers, in particular to a model confidence evaluation method and device, equipment, a medium and a program product, and the method comprises the steps: obtaining an information evaluation set; inputting test questions in the information evaluation set into the first model and the second model respectively to obtain to-be-evaluated answers output by the first model and reference answers output by the second model; inputting the reference answers and standard answers of the test questions in the information evaluation set into a pre-trained evaluation model, and outputting first deviation information between the reference answers and the standard answers; when determining that the test question and the standard answer meet a preset confidence condition based on the first deviation information, inputting the to-be-evaluated answer and the standard answer into an evaluation model, and outputting second deviation information between the to-be-evaluated answer and the standard answer; generating a confidence evaluation report of the first model according to the second deviation information; according to the scheme, the confidence test of the test questions and the standard answers is realized, and the universality and the mobility of the evaluation method are improved through the arrangement of the evaluation model.

The invention provides an AI-based talent whole-process management method and device, electronic equipment and a medium, and the method comprises the steps: obtaining a recruitment application submitted by a recruiter, building a recruitment post after the recruitment application passes the approval, and generating the post responsibility information of the recruitment post based on a first sub-module of an AI model; obtaining resume information of the applicant, executing a resume information evaluation task based on a second sub-module of the AI model, and calculating a matching degree score between the resume information and the recruitment post; an interview invitation is sent to the applicant based on the matching degree score, an interview evaluation task is executed through a third sub-module of the AI model, the applicant receiving the interview invitation is interviewed, and an evaluation result of the interviewed applicant is obtained; and determining candidates based on the evaluation result, and sending a review notification or an in-job invitation notification to the candidates. According to the invention, the recruitment process is optimized, and the recruitment quality is improved.

The embodiment of the invention provides an evaluation report generation method and device, electronic equipment and a storage medium, and belongs to the technical field of information processing. The method comprises the following steps: acquiring a plurality of information evaluation items; performing form design on all the information evaluation items to obtain evaluation items; integrating all the evaluation items to generate an evaluation report template; receiving a filling request and filling content sent by a user, and generating a request number; determining a target project identifier, and filling the filling content into the target evaluation project according to the target project identifier; in response to a submission instruction sent by a user, content auditing is performed on the updated evaluation report template based on the request number; when the updated evaluation report template does not pass content auditing, determining failed evaluation items which do not pass content auditing, and generating auditing suggestions; and filling the evaluation report template with the auditing opinions to obtain a feedback evaluation report. The embodiment of the invention can feed back the auditing result of the user in time and ensure the timeliness of evaluation.

The invention relates to the technical field of information fusion, in particular to a power equipment state information fusion system and method. The method comprises the following steps: obtaining time sequence signal data and infrared thermal image data, and constructing a power equipment heterogeneous fusion data set; acquiring power equipment position and power equipment environment data, performing multi-layer cross-mode data coupling in combination with the power equipment heterogeneous fusion data set, and constructing a power equipment information evaluation model based on the multi-layer cross-mode data coupling; therefore, through multi-source data fusion and cross-level evaluation model construction, the problems of a traditional power equipment monitoring system in the aspects of information island and prediction accuracy are solved, and the health management capability and fault early warning efficiency of power equipment are improved.

The invention belongs to the field of artificial intelligence and knowledge graph application, and relates to an intent-driven large model knowledge graph intelligent question and answer method and an intelligent agent. The method comprises the steps of receiving a question input by a user, and identifying a potential intention of the user; according to the identified intention and the information stored in the shared memory, selecting a proper tool from a tool only based on a large language model, a tool based on semantic analysis, a tool based on sub-graph retrieval and a general tool for reasoning; new information generated by the tool is updated into the shared memory; and in each iteration, the large language model evaluates the current reasoning state according to the information stored in the shared memory, and triggers a tool with a corresponding intention to perform supplementary reasoning according to the current reasoning state until enough information is collected to draw a conclusion or a preset upper limit of the number of iterations is reached. The method can improve the reasoning efficiency and accuracy, can be applied to complex question answering in various fields, and provides high-quality question answering service for users.

The invention discloses an occupational information evaluation method based on rule judgment and a large language model, and belongs to the technical field of information processing, and the method comprises the following steps: S1, obtaining a multi-modal document uploaded by a user, converting the multi-modal document into a Markdown format, and completing a shunting operation and a converging operation; s2, extracting a named entity of the multi-modal document after the confluence operation, and constructing a knowledge graph; s3, obtaining query content, and matching most relevant regulation content by using the knowledge graph; s4, generating specific improvement suggestions by utilizing the rule model according to the most relevant regulation content; and S5, adjusting the specific improvement suggestions by using a large language model to obtain a final evaluation result. Through the OCR technology, regular expression analysis and context-associated picture analysis, unified structured processing of texts, tables and pictures is achieved, and the problem that unstructured data are difficult to integrate in a traditional method is effectively solved.

The invention relates to the technical field of carbon emission measurement and calculation, in particular to a pavement construction process carbon emission prediction system and method based on intelligent monitoring, and the system comprises a data collection module which monitors and collects the energy consumption data in the pavement construction process in real time; the data analysis module is used for determining a corresponding carbon emission factor according to the energy consumption data and determining the carbon emission in the pavement construction process according to the energy consumption data and the carbon emission factor; the information evaluation module receives the carbon emission in the pavement construction process, compares the carbon emission with a set threshold value, and adjusts a construction scheme in real time according to a comparison result; and the control center is connected with the data acquisition module, the data analysis module and the information evaluation module, realizes information transmission, and intensively displays output information of the data acquisition module, the data analysis module and the information evaluation module. Real-time monitoring of energy consumption data in the pavement construction process is achieved, decisions and on-site adjustment suggestions are given for a construction scheme, and carbon emission in the pavement construction process is reduced.

The invention provides a multi-modal online car-hailing comfort evaluation network model optimized by a genetic algorithm and a construction method of the multi-modal online car-hailing comfort evaluation network model. Constructing an end-to-end vision-odor multi-modal comfort evaluation network model, wherein the network model is composed of tidiness feature extraction, odor pseudo image feature extraction and multi-modal information evaluation; a genetic algorithm is combined with a back propagation learning method, and key parameters of the multi-modal evaluation neural network model are optimized; and a trained end-to-end multi-modal evaluation network model is used to carry out multi-dimensional quantitative evaluation on the comfort level of the in-vehicle environment. According to the method, odor pseudo image-tidiness image features can be effectively fused; compared with the method for evaluating the comfort level of the environment in the vehicle from the perspective of peculiar smell or cleanliness, the method has the advantages that the accuracy of the evaluation result after the cleanliness-peculiar smell information is effectively fused is obviously improved, and the evaluation result has higher robustness and reference value.

The invention discloses an insurance evaluation large model training method and device, equipment and a storage medium, and relates to the technical field of data processing. Comprising the steps of performing data compliance processing on medical training data and financial training data provided by participants participating in model training through a federal learning architecture in a preset distributed data management component to obtain compliance data; reading compliance data from a preset distributed data management component based on a preset data fusion engine, and performing fusion processing to obtain cross-domain fusion data; obtaining cross-domain fusion data by using a preset large model joint training framework, and generating a target model training task based on the model training capability corresponding to each participant; and training the to-be-trained model according to the target model training task and the cross-domain fusion data to obtain a target insurance evaluation large model so as to perform medical insurance information evaluation based on the target insurance evaluation large model. In this way, efficient retrieval and compliance sharing of medical and financial cross-domain data resources can be achieved.

The invention relates to the field of mine geological environment monitoring and evaluation, in particular to an intelligent mine geological environment information evaluation method based on a knowledge graph. The method comprises the steps that mine geological environment data are collected and preprocessed, a knowledge graph is constructed based on the preprocessed mine geological environment data, and embedded vectors of nodes in the knowledge graph are obtained; calculating space-time similarity among the nodes based on the embedded vectors of the nodes; iteratively optimizing the embedded vectors of the nodes by adopting a graph neural network to obtain final embedded vectors of the nodes; a self-adaptive edge weight updating mechanism is introduced, and edge weights among the nodes are dynamically adjusted to obtain final edge weights among the nodes; and comprehensively evaluating the mine geological environment based on the final embedding vector of the nodes and the final edge weight between the nodes. The problems that a traditional monitoring and analysis method neglects spatial and temporal characteristics of a mine geological environment, and accurate prediction and risk early warning are difficult to carry out are solved; and real-time self-adaptive updating cannot be carried out, so that the timeliness of an evaluation result is poor.

The invention relates to the technical field of on-site safety management, and discloses a construction site on-site safety inspection system and method based on big data, and the system comprises a construction area construction information collection module, a construction area construction information evaluation module, and a construction area construction danger feedback module. Construction danger type intelligent analysis is carried out by storing different types of construction site danger image parameters based on big data in combination with an intelligent recognition algorithm and real-time inspection position image parameters of a current construction area, and accurate monitoring of the construction danger type of the on-site inspection position of the whole construction site is achieved. Construction site safety inspection area space position coordinate parameters are autonomously set, and coordinate data matching is carried out on the construction site safety inspection area space position coordinate parameters and current construction area real-time inspection position image parameters in combination with an intelligent search algorithm, so that a construction site safety inspection area where current construction dangers are located is accurately identified; partitioned, efficient and accurate positioning of the construction danger occurrence site of the construction site is achieved.

The present invention discloses a multi-dimensional information evaluation method and system for an enterprise security index, and relates to the technical field of data analysis. The present invention retrieves business data from an enterprise database to obtain current real-time business project generation data; constructs a business project data set in combination with data normalization processing; performs business feature transformation based on the business project data set, performs development feature value analysis on the business project data, and constructs a business project development feature curve; retrieves business project data feature prediction curve data, performs comparative analysis in combination with the current real-time business project development feature curve, and performs abnormal feature point judgment and classification; performs abnormal risk analysis on local adjacent feature points in combination with feature point classification results; performs risk assessment on the current business project in combination with the abnormal risk analysis data of each local adjacent feature point; visually outputs the risk analysis data of the corresponding business project of the enterprise, and provides comprehensive feedback on the risk assessment results.

The invention relates to the technical field of medical signal processing, and discloses an anesthesia depth multi-level classification information assessment method and system, and the method comprises the steps: collecting an electroencephalogram signal, and carrying out the preprocessing operation of the electroencephalogram signal, and obtaining signals of five frequency bands; global field power is calculated for all electroencephalogram signals of each frequency band, and a peak topographic map is identified; performing clustering analysis to extract a micro-state template, and matching and reconstructing a micro-state time sequence; and establishing a feature screening and classification identification model, inputting each frequency band signal into the feature screening and classification identification model, screening the extracted micro-state features, carrying out multi-level classification on the anesthesia depth, and finally outputting an anesthesia depth evaluation result. The system comprises an electroencephalogram collecting and processing module, a clustering reconstruction module and an anesthesia depth evaluation module. According to the method, the preprocessed electroencephalogram signals are decomposed through Hilbert-Huang transform, and anesthesia depth evaluation is achieved.

The invention discloses an intelligent auxiliary analysis method and system for a distribution network, and relates to the technical field of power distribution networks, and the method comprises the following steps: collecting alarm information data and operation data of each monitoring point of the power distribution network in real time; identifying the alarm information data based on a time window mechanism and filtering the same alarm information repeatedly appearing in the G time; constructing an alarm information evaluation model for the filtered alarm information data and operation data based on a BP neural network to obtain an alarm information evaluation coefficient; acquiring alarm information evaluation coefficients around each monitoring point of the power distribution network in the same time interval, and constructing a first data set; and according to the average value of the first data set of each monitoring point, performing priority ranking on the alarm events, and preferentially processing the monitoring point area with the maximum average value. The method comprehensively considers the evaluation mode of various factors, enables the alarm information evaluation to be more comprehensive and accurate, and provides a more reliable decision basis for operation and maintenance personnel.

The invention discloses a motor multi-parameter identification method based on sampling noise excitation and bias compensation, and the method comprises the steps: firstly obtaining the voltage and sampling current of a motor, and enabling the sampling current to comprise sampling noise; then constructing an information evaluation index, and adjusting a proportional parameter of a current controller to improve the identification information amount; then, shaft voltage feed-forward compensation is constructed to suppress electromagnetic torque pulsation; recursion is carried out by adopting a least square method based on the discrete linear parameterization model to obtain a parameter initial value and a covariance matrix; secondly, estimating offset-containing parameters on line, and calculating a sampling noise variance by using an estimated residual error; and finally, obtaining unbiased parameter estimation according to bias compensation iteration updating so as to output online estimation values of stator resistance, direct-axis inductance Ld, quadrature-axis inductance Lq and permanent magnet flux linkage.

The invention provides a prompt information evaluation method and device, electronic equipment, a medium and a program product, and can be applied to the technical field of artificial intelligence. The method comprises the steps that a pre-constructed query statement test data set is obtained, the query statement test data set at least comprises a plurality of error query statements, and each error query statement is used for triggering a target database to generate error prompt information in the execution process; inputting the error query statement into the target database, obtaining return information of the target database, and generating error prompt information based on the return information; based on the error prompt information, utilizing a large language model to obtain at least one corresponding correction query statement; and verifying the corrected query statement in the target database, and evaluating the quality of the error prompt information based on a verification result.

1. Name of this design product: Graphical user interface for a central control visualization platform for electronic equipment in a civil aviation cabin. 2. Purpose of this design product: an electronic device. 3. The key design point of this design product lies in the graphical user interface on the screen. 4. The picture or photo that best illustrates the key points of the design: main view. 5. The physical product is of conventional design and other views are omitted. 6.Purpose of the graphical user interface: used to summarize and update data display of civil aviation cabin information. The interface displayed in the main view is the graphical user interface of this software. The middle area of ​​the main view shows the flight cabin data summary, and the left and right sides show data indicators. The left side shows the cabin central control visualization platform, cabin service management and power consumption trend respectively. The right side shows the cabin response information evaluation, the real-time status of temperature and humidity in each cabin area and the cabin service demand data statistics. The lower middle area shows the real-time video inside or outside the cabin. In the main view, click on the "In-cabin response information evaluation" text area, and a pop-up window will display data indicators such as in-cabin service evaluation and in-cabin response times.

The embodiments of the present disclosure relate to an evaluation model construction method, a model generation information evaluation method, and an apparatus, said construction method comprising: acquiring a plurality of multimedia samples generated by a target generation model in respect of a prompt information sample, and acquiring evaluation result labels respectively corresponding to the plurality of multimedia samples in various preset evaluation dimensions; by means of an information processing network, separately fusing the prompt information sample with target multimedia samples, so as to obtain multiple pieces of sample fusion information; on the basis of the multiple pieces of sample fusion information and by means of a plurality of evaluation networks to be trained, generating network evaluation results respectively corresponding to the plurality of multimedia samples in the various evaluation dimensions; on the basis of the evaluation result labels and the network evaluation results respectively corresponding to the plurality of multimedia samples in the various evaluation dimensions, training the plurality of evaluation networks; and, on the basis of the information processing network and the plurality of trained evaluation networks, constructing an evaluation model. Thus, the embodiments of the present disclosure effectively ensure the reliability of evaluating model generation information.

The invention discloses a reverse guidance type intelligent interaction method and system for realizing dynamic information assessment. The system comprises an information sufficiency assessor, an information sufficiency assessment device and a dynamic information assessment device, wherein the information sufficiency assessor is used for dynamically and quantitatively assessing and identifying key information gaps on the basis of an information entropy theory; the reverse questioning strategy generator is used for generating an optimal reverse guidance questioning sequence for supplementing a key information gap based on a Markov decision process model; the hierarchical problem decomposition engine is used for recursively decomposing a complex problem into a plurality of sub-problems by adopting a semantic tree decomposition technology, and fusing solution results of the sub-problems by utilizing an evidence theory fusion mechanism to form a problem solution; the personalized solution synthesizer is used for analyzing the cognitive level, learning style and attention characteristics of the user portrait, performing matching optimization on a problem solution and the user portrait, and calculating and measuring the semantic matching degree of the problem solution and the user portrait; and applying reinforcement learning to continuously optimize an output strategy according to user feedback to generate a personalized solution.

The invention relates to the technical field of medicine, and provides a medicine clinical value evaluation method and system, electronic equipment and a storage medium, and the system comprises a medicine information collection module which is used for obtaining first information of a to-be-evaluated target medicine; the evaluation standard construction module is used for setting an evaluation standard based on a preset evaluation requirement; the standard quantitative assignment module is used for performing weight assignment on the evaluation standard; the evaluation evidence integration module is used for acquiring second information of the target drug based on the evaluation standard set by the evaluation standard construction module and the first information; and the drug value evaluation module is used for evaluating the clinical value of the target drug based on the second information of the target drug by applying an evaluation standard. The method and the device are used for overcoming the defects of dimensionality stiffness, quantization distortion, evidence dispersion and decision lag during evaluation of the clinical value of the medicine in related technologies. The method and the device can be used for realizing scientific evaluation, dynamic management and scenarized decision of the clinical value of the medicine.

The invention discloses an enterprise credit information assessment system and method, and the system comprises a data collection module, a data screening module, a text recognition module and a risk assessment module: the data collection module is used for collecting enterprise public information corresponding to a to-be-assessed enterprise; the data screening module is used for cleaning and screening the enterprise public information according to the consistency among various information in the enterprise public information to obtain to-be-evaluated information; the text recognition module is used for performing text recognition on the to-be-assessed information and extracting multiple pieces of target information corresponding to the to-be-assessed enterprise in the to-be-assessed information; and the risk assessment module is used for analyzing and predicting each piece of target information based on a preset identification model, determining the credit risk of the to-be-assessed enterprise, generating an enterprise credit information report based on the credit risk, and completing credit information assessment of the to-be-assessed enterprise. According to the invention, accurate and efficient evaluation of enterprise credit information is realized.

The invention discloses a public health real information evaluation and early warning system based on an AI technology, and relates to the field of artificial intelligence technology application, and the system comprises a database module which is used for storing information collected by a collection module and providing powerful support for subsequent model construction and real-time response, and a data collection module which is used for collecting real-time information of public health. The invention discloses a public health data collection and processing method, which is used for collecting data related to public health from various sources (such as news, social media, professional forums and the like), including multi-modal data such as texts, pictures, videos and the like, and ensuring that the data sources cover data sources of a knowledge graph and data sources of a false information base. The information processing efficiency is greatly improved through a knowledge graph and advanced model construction, the system has strong adaptive capacity for novel false information, the system can rapidly learn and recognize a new false information mode through fusion of the knowledge graph and a large model, and the risk caused by the false information is effectively reduced.

The invention discloses a public praise information evaluation method based on data analysis, and relates to the technical field of information evaluation. Content generated in a forum is acquired through a crawler technology; putting the original sentences into a filter, and obtaining official public praise aspects after screening of the filter; extracting adjective words, degree adverbs and negative words in an original sentence through an NLP technology, and evaluating the adjective words, the degree adverbs and the negative words; combining the emotion assignment of the degree adverb and the adjective to obtain a word-of-mouth evaluation positive degree score, and judging the word-of-mouth evaluation positive degree according to a scoring result; outputting the obtained public praise polarity and public praise evaluation positive degree; according to the method, the whole word-of-mouth mining task is divided into a plurality of sub-tasks to be split and simplified, the problems of data scarcity and marking quality are relieved to a great extent, and by calculating the word-of-mouth positive degree score, the positive degree of word-of-mouth evaluation can be judged while high accuracy is kept when more word-of-mouth information is extracted.

The invention discloses a highway information evaluation method based on data management, and relates to the technical field of highway evaluation, and the method comprises the evaluation steps: S1, collecting highway traffic flow data, pavement state data and local meteorological data, and carrying out the preprocessing of the collected data; s2, extracting a correlation calculation characteristic value from the preprocessed data, synchronizing the processed data according to a time dimension, and enabling the collected data to correspond to each other; and S3, calculating a pavement damage rate, a traffic jam rate and a meteorological influence rate based on the data. According to the invention, the evaluation score is obtained by comprehensively considering each factor, a scientific basis is provided for the management decision, the grading and maintenance decision can determine the condition grade, make the maintenance strategy according to the grade, realize reasonable resource allocation, improve the overall operation and maintenance management level of the expressway, and avoid the problem of dependence on single data evaluation. The processing flow is effectively improved, and accurate evaluation and analysis are carried out.

The invention discloses a method for positioning and detecting key parts of hatching eggs. The method comprises the following steps: acquiring transmission images in hatching eggs of different days through a hatching egg image acquisition device, and training and establishing a hatching egg physiological state judgment model and a hatching egg key part segmentation model; the discrimination model outputs a transmission image of the fertilized egg to be positioned and detected, the key part is identified after the transmission image is processed by the segmentation model, then the optimal coordinate point extracted by the allantoic fluid is positioned by adopting an image processing method, and the allantoic fluid is extracted to realize the detection of the hatching egg. According to the method, a dual model of hatching egg development evaluation and key part identification is established, interference of human subjective factors is weakened, and detection and classification errors are reduced; the allantoic fluid coordinate positioning method reduces the risk of puncturing and damaging the embryo and improves the allantoic fluid extraction success rate; according to the invention, rapid and accurate hatching egg information evaluation and male and female chick embryo difference characteristic detection can be provided according to hatching application requirements.

The invention discloses a multi-modal-based commodity pushing method and system, and relates to the technical field of commodity pushing, and the method specifically comprises the following steps: analyzing collected multi-modal feature information, and generating a recommendation correlation coefficient and a compression loss influence index of each piece of feature information in the multi-modal feature information; constructing an important information evaluation model for the recommendation correlation coefficient and the compression loss influence index of each piece of generated feature information, respectively generating an important coefficient of each piece of feature information, and evaluating the importance degree of each piece of feature information in the multi-modal feature information to a recommendation result; dividing each piece of feature information into high-importance feature information, medium-importance feature information and low-importance feature information according to an evaluation result; and respectively taking corresponding data compression measures on the feature information with different importance degrees based on a classification result of each piece of feature information. According to the method, the problem of priority retention of important features in multi-modal data compression is solved, and the recommendation precision is optimized.