Learning material retrieval evaluation autonomous optimization system and method

Abstract

The invention discloses a learning material retrieval evaluation autonomous optimization system and method, and belongs to the technical field of learning material retrieval, and the method comprises the steps: determining each target resource channel, and setting a target resource channel account of each target resource channel; docking each search channel; receiving a retrieval condition of a user, retrieving the retrieval condition through each search channel to obtain each piece of initial retrieval information, and performing initial screening on each piece of initial retrieval information to obtain each piece of to-be-selected retrieval information; evaluating each piece of to-be-selected retrieval information to obtain each retrieval recommendation result, and displaying each retrieval recommendation result to the user in columns according to whether the user is charged or not; classifying the retrieval behavior data according to the retrieval analysis data to obtain optimized classifications; evaluating each optimization classification, and judging whether the optimization classification meets a retrieval standard or not; marking the optimization classifications which do not meet the retrieval standard as target optimization classifications; and performing retrieval optimization according to the target optimization classification.

Description

Technical Field

[0001] The present invention belongs to the technical field of learning material retrieval, and specifically relates to a learning material retrieval evaluation autonomous optimization system and method. Background Art

[0002] In the current era of digital learning, the acquisition of learning materials is crucial to learners' learning outcomes. However, existing learning material retrieval systems generally face a core problem: the learning materials in the built-in database are often not comprehensive enough to meet the diverse learning needs of users. These systems usually rely on pre-built databases to store and retrieve learning materials, but due to the complexity and cost of data collection, organization and maintenance, the content of the database is often limited and cannot cover all fields and the latest knowledge.

[0003] In addition, with the rapid development of the Internet, the learning materials on the Internet are updated very quickly, especially news, current affairs commentary and other content that need to be updated in real time. However, due to technical limitations or insufficient resource investment, the existing learning material retrieval system is often unable to capture and integrate these latest learning materials into its built-in database in a timely manner. As a result, when searching, users may only be able to obtain outdated or no longer applicable materials, which seriously affects learning efficiency and effectiveness.

[0004] In response to the above problems, although some systems try to keep the data current by regularly updating the database, this method is not only inefficient, but also difficult to ensure that all data can be updated in a timely manner. Especially when faced with learning materials that need to be updated in real time, the traditional database update mechanism seems to be unable to cope with it.

[0005] Based on this, the present invention provides a learning material retrieval and evaluation autonomous optimization system and method. Summary of the invention

[0006] In order to solve the problems existing in the above scheme, the present invention provides a learning material retrieval evaluation autonomous optimization system and method.

[0007] The purpose of the present invention can be achieved through the following technical solutions:

[0008] A learning material retrieval evaluation autonomous optimization system, comprising an account management module, a retrieval module and an evaluation optimization module;

[0009] The account management module is used to manage channel accounts, summarize various learning needs and the demand proportions corresponding to each learning need, and simulate and set a simulated search set according to each learning need;

[0010] Perform search simulation based on the simulated search set to obtain various simulated search results; analyze each of the simulated search results to determine each target resource channel; and set a target resource channel account according to each of the target resource channels.

[0011] Furthermore, the target resource channel account is dynamically updated.

[0012] Furthermore, the method for analyzing each simulated search result includes:

[0013] Obtain each search recommendation result corresponding to each simulated search data; identify the source channel and source channel information corresponding to each search recommendation result;

[0014] Classify each search recommendation result according to each source channel to obtain channel classification data corresponding to each source channel; and mark the corresponding demand proportion for each search recommendation result in each channel classification data;

[0015] Identify the union part between the classified data of each channel, and identify the independent part in the classified data of each channel according to each of the union parts; mark each of the source channels with the independent part as the target resource channel;

[0016] The remaining source channels are screened according to the respective union parts to obtain the respective target resource channels.

[0017] Furthermore, the method of screening the remaining source channels according to the combined parts includes:

[0018] Mark each search recommendation result corresponding to the source channel as i, where i=1, 2, ..., n, and n is a positive integer; perform access evaluation on each search recommendation result according to the source channel information to obtain a corresponding access value; mark the access value as FEi;

[0019] According to the formula Calculate the corresponding screening value;

[0020] Where: PA is the screening value; βi is the demand proportion corresponding to the corresponding search recommendation result;

[0021] The source channels are screened according to the screening values ​​to obtain the target resource channels.

[0022] The retrieval module is used by users to retrieve learning materials, and includes an input unit and a retrieval unit;

[0023] The input unit is used for the user to input the search condition and send the search condition to the search unit;

[0024] The retrieval unit is used to retrieve learning materials based on the retrieval information; connect to various search channels; receive the user's search conditions, search the search conditions through various search channels, obtain initial search information, perform initial screening on each of the initial search information, and obtain each candidate search information; evaluate each candidate search information, obtain each search recommendation result, and display each search recommendation result to the user in columns according to whether it is charged.

[0025] Furthermore, the method for initially screening each initial search information includes:

[0026] An invalid evaluation model is established, and the expression of the invalid evaluation model is:

[0027]

[0028] Where: s is the input data, the input data is the search information and search conditions; the output data is the invalid evaluation value WP(s);

[0029] Analyze each of the initial search information through an invalid evaluation model to obtain an invalid evaluation value corresponding to each of the initial search information; and remove each of the initial search information with an invalid evaluation value of 0;

[0030] The remaining initial search information is identified, and the same initial search information is classified into one category to obtain each search information category; the initial search information in each search information category is evaluated and screened to obtain a candidate search information corresponding to each search information category.

[0031] Furthermore, the method for evaluating and screening each initial search information in the search information classification includes:

[0032] Estimate the access value corresponding to each initial search information; identify the target resource channel corresponding to each initial search information; evaluate each target resource channel according to the preset additional evaluation items to obtain the corresponding additional evaluation value;

[0033] Calculate the corresponding screening value according to the formula PB = b1 × FZ-b2 × FE;

[0034] Where: PB is the screening value; FZ is the additional value; FE is the access value; b1 and b2 are both proportional coefficients, with a value range of 0≤b1<1, 0 <b2≤1;

[0035] The initial search information with the largest screening value is selected as the search information to be selected.

[0036] The evaluation and optimization module is used to perform search optimization based on the search behavior data of each user, obtain the search behavior data of each user and the corresponding search analysis data, and the search analysis data includes the search conditions and the search recommendation results;

[0037] Classify each search behavior data according to each search analysis data to obtain each optimized classification;

[0038] Evaluate each optimization classification to determine whether it meets the search criteria; mark the optimization classifications that do not meet the search criteria as target optimization classifications; and perform search optimization based on the target optimization classifications.

[0039] Furthermore, the method of classifying each retrieval behavior data according to each retrieval analysis data includes:

[0040] Establish a retrieval equivalence judgment model, the expression of the retrieval equivalence judgment model is:

[0041]

[0042] In the formula: q is the input data, the input data is the two search analysis data to be compared; the output data is the search equivalent value YP(q);

[0043] The retrieval analysis data are analyzed by the retrieval equivalence judgment model to obtain the retrieval equivalence values ​​between the retrieval analysis data; the retrieval behavior data are classified according to the retrieval equivalence values ​​to obtain the optimized classifications.

[0044] Furthermore, the method for evaluating each optimization classification includes:

[0045] Identify each search behavior data corresponding to the optimization classification, and obtain the corresponding search duration according to each search behavior data;

[0046] A behavior evaluation model is established. The expression of the behavior evaluation model is:

[0047]

[0048] Where: pj is the input data, the input data is the retrieval behavior data, j = 1, 2, ..., m, m is a positive integer; T is the retrieval time; T0 is the preset time; the output data is the behavior evaluation value XQ (pj);

[0049] Analyze each retrieval behavior data in the optimized classification through the behavior evaluation model to obtain the corresponding behavior evaluation value;

[0050] According to the formula Calculate the corresponding optimization evaluation value;

[0051] Where: PC is the optimized evaluation value;

[0052] The optimization classification whose optimization evaluation value is lower than the threshold X2 is marked as the target optimization classification.

[0053] A learning material retrieval evaluation autonomous optimization method, the method comprising:

[0054] Summarize various learning needs and the demand proportions corresponding to each learning need, and simulate and set up simulated retrieval sets according to each learning need;

[0055] Perform search simulation based on the simulated search set to obtain each simulated search result; analyze each simulated search result to determine each target resource channel; set the target resource channel account according to each target resource channel;

[0056] Connecting with various search channels; receiving the user's search conditions, searching the search conditions through various search channels, obtaining various initial search information, initially screening various initial search information, obtaining various candidate search information; evaluating various candidate search information, obtaining various search recommendation results, and displaying various search recommendation results to the user in columns according to whether they are charged;

[0057] Classify each search behavior data according to each search analysis data to obtain each optimized classification; the search analysis data includes search conditions and each search recommendation result;

[0058] Evaluate each optimization classification to determine whether it meets the search criteria; mark the optimization classifications that do not meet the search criteria as target optimization classifications; and perform search optimization based on the target optimization classifications.

[0059] Compared with the prior art, the present invention has the following beneficial effects:

[0060] Through the cooperation among the account management module, the search module and the evaluation and optimization module, the real-time access and search of the massive learning materials on the Internet are realized, which greatly expands the source range of learning materials and effectively solves the problem of incomplete learning materials in the existing system's built-in database. At the same time, because it can capture the latest information on the Internet in real time, especially news, current affairs commentary and other materials that need to be updated quickly, it ensures that the materials retrieved by users are the latest and most accurate, and improves the timeliness of learning materials.

[0061] Since this solution is based on a network search engine, it has good flexibility and scalability. With the continuous development and improvement of search engine technology, the retrieval platform proposed by the present invention can be continuously upgraded and optimized to meet the needs of more extensive and complex learning material retrieval. BRIEF DESCRIPTION OF THE DRAWINGS

[0062] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. 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 creative work.

[0063] Figure 1 This is a principle block diagram of the present invention. DETAILED DESCRIPTION

[0064] The technical solution of the present invention will be clearly and completely described below in conjunction with the embodiments. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0065] like Figure 1 As shown, a learning material retrieval evaluation autonomous optimization system includes an account management module, a retrieval module and an evaluation optimization module;

[0066] The account management module is used to manage channel accounts, determine the various possible learning needs and the demand proportion of each learning need according to the service scope of the platform, and simulate and set a simulated retrieval set according to each learning need, that is, set the detection input data set corresponding to the retrieval needs that subsequent users may search according to the learning needs; search for information on the Internet through the simulated retrieval set to obtain various simulated retrieval results; it means to search according to the subsequent retrieval method of the present invention; analyze each simulated retrieval result to determine each target resource channel; set the target resource channel account according to each target resource channel; that is, the corresponding target resource channel can be accessed according to the target resource channel account, because the current part of the learning materials are in each independent resource channel, and some are exclusive copyrights, so the above-mentioned learning materials need to be logged in to the corresponding target resource channel to access and obtain; the setting of the specific target resource channel account can be negotiated and handled by the platform and the management personnel of the corresponding target resource channel, or it can be established by itself, but it is generally negotiated.

[0067] In one embodiment, the target resource channel is in a dynamic update state, that is, the learning requirements and the simulated search set are subsequently updated according to the actual user search records, thereby achieving the update of the target resource channel.

[0068] In one embodiment, the target resource channel may be determined based on existing technology or manually.

[0069] In one embodiment, the method for analyzing each simulated search result includes:

[0070] Identify the source channel and source channel information corresponding to each search recommendation result according to each simulated search data; the source channel information is the relevant information of the channel, such as name, access conditions, charging standards and other relevant data;

[0071] Identify each source channel, classify each search recommendation result according to the source channel, and obtain channel classification data corresponding to each source channel; and mark the corresponding demand proportion of each search recommendation result in the channel classification data, that is, the learning proportion of the corresponding learning demand;

[0072] Identify the union part between the classified data of each channel, because a simulated search data may have multiple search recommendation results, some of which are the same, but only have different source channels. Therefore, the channel classification data corresponding to multiple source channels may have a union intersection, and then determine the union part, that is, the set corresponding to the common search recommendation results;

[0073] Identify the independent parts in the classified data of each channel according to the parts of each union. The independent parts refer to the search recommendation results corresponding to them that are not included in the classified data of other channels. Mark the source channels with independent parts as target resource channels.

[0074] The remaining source channels are screened according to the combined parts to obtain the target resource channels.

[0075] In one embodiment, the method of screening the remaining source channels according to the combined parts includes:

[0076] Each search recommendation result corresponding to the source channel is marked with i, where i=1, 2, ..., n, where n is a positive integer; each search recommendation result is evaluated for access and a corresponding access value is obtained. The access value is set according to whether the search recommendation result needs to be paid. If no payment is required, the access value is 0. If payment is made separately, the corresponding price is directly used as the access value. If access is made according to membership and other conditions, the average price is calculated based on the membership and other condition prices and the number of corresponding search recommendation results. The average price can be directly used as the access value, or the average price can be amplified as the access value. The specific setting is based on the needs of the platform.

[0077] The access value is marked as FEi; according to the formula Calculate the corresponding screening value;

[0078] Where: PA is the screening value; βi is the demand proportion corresponding to the corresponding search recommendation result;

[0079] The source channels with screening values ​​lower than the threshold X1 are eliminated, and the remaining source channels are marked as target resource channels.

[0080] The retrieval module is used by users to retrieve learning materials, and includes an input unit and a retrieval unit;

[0081] The input unit is used for users to input search conditions, such as keywords, etc.; it is the initial interface displayed by the search module to the user; when the user needs to search for learning materials, the corresponding search conditions are input through the input unit.

[0082] The retrieval unit is used to retrieve learning materials according to the retrieval information; connect to various existing search engines, such as Baidu, 360 browser, etc.; mark as search channels; identify the retrieval conditions entered by the user in the input unit, retrieve the retrieval conditions through various search channels, and obtain various initial retrieval information, which is each retrieval information retrieved by each search channel; collect in combination with the corresponding target resource channel account;

[0083] Performing initial screening on each initial search information to obtain each candidate search information;

[0084] Each candidate search information is evaluated to obtain each search recommendation result, and each search recommendation result is displayed to the user in columns according to whether it is charged or not.

[0085] For paid search recommendation results, not all results can be displayed. Only the visible part can be displayed, and subsequent access will be based on the platform's service policy.

[0086] In one embodiment, after each candidate search information is determined, each search recommendation result can be determined using an existing search recommendation algorithm, such as based on a matching degree with the search condition.

[0087] Methods for initially screening each initial search information include:

[0088] An invalid evaluation model is established. The invalid evaluation model is used to identify advertising information, non-learning materials and other information that are not related to learning materials in the retrieval information, and regard them as invalid information. Invalid information is defined based on the existing historical retrieval data, and then the retrieval information can be identified and judged based on the corresponding invalid information definition to determine whether it is invalid information. However, in actual applications, in order to improve the recognition accuracy, effective information is generally used as a benchmark for recognition, which is convenient for training and reduces the amount of data processing. For example, based on the correlation and matching between the retrieval conditions and each retrieval information, the existing technology can be used for judgment, and non-target resource channels can be regarded as invalid information. If the requirements are met, it is regarded as valid data. The expression of the invalid evaluation model is: Where: s is the input data, and the input data is the retrieval information and retrieval conditions; the output data is the invalid evaluation value WP(s).

[0089] Analyze each initial retrieval information through the invalid evaluation model to obtain the invalid evaluation value corresponding to each initial retrieval information; eliminate the initial retrieval information with an invalid evaluation value of 0; in other embodiments, it is also possible to eliminate the invalid initial retrieval information based on the prior art.

[0090] Identify the remaining initial retrieval information, classify the same initial retrieval information into one category to obtain each retrieval information classification; evaluate and screen each initial retrieval information in the retrieval information classification to obtain a candidate retrieval information corresponding to each retrieval information classification; that is, only one of the duplicate initial retrieval information is retained.

[0091] The method for evaluating and screening each initial retrieval information in the retrieval information classification includes:

[0092] Estimate the access value corresponding to each initial retrieval information; identify the target resource channels corresponding to each initial retrieval information, set additional evaluation items according to the needs of the platform party, and evaluate each target resource channel based on the additional evaluation items to obtain the corresponding additional evaluation value; the additional evaluation items are used to evaluate the target resource channels corresponding to the initial retrieval information, such as evaluating from relevant factors such as browsing process items, in-channel advertising conditions, and security, that is, evaluating from the advantages and disadvantages of each channel, and using the existing channel evaluation method for evaluation. Exemplarily, establish a corresponding channel evaluation model based on neural networks such as CNN network or DNN network, and train it through an artificial method to establish a corresponding training set. The training set includes input data and output data. The input data is the relevant information of the target resource channel, and the output data is the additional evaluation value; evaluate through the successfully trained channel evaluation model; in other embodiments, the platform party can also manually evaluate each target resource channel and set the corresponding additional evaluation value.

[0093] Calculate the corresponding screening value according to the formula PB = b1×FZ - b2×FE;

[0094] Where: PB is the screening value; FZ is the additional value; FE is the access value; b1 and b2 are both proportionality coefficients, and the value range is 0 ≤ b1 < 1, 0 < b2 ≤ 1, that is, the additional evaluation items can be not considered when needed.

[0095] Select the initial retrieval information with the largest screening value as the candidate retrieval information.

[0096] The evaluation and optimization module is used to perform search optimization according to the search behavior data of each user, and the search behavior data is from the display of the search recommendation results to the user to the user determining the required learning materials or not retrieving the required learning materials; illustratively, a result submission function can be set, that is, if the user does not retrieve the required search materials, he can click the displayed result submission button to indicate that the required learning materials have not been retrieved; if the user does not click the result submission button, it is deemed that the required learning materials have been retrieved; whether the learning materials have been retrieved can be determined by other means;

[0097] Identify each user's search behavior data and corresponding search analysis data, the search analysis data including search conditions and each search recommendation result;

[0098] According to each retrieval analysis data, each retrieval behavior data is classified, that is, each retrieval behavior data with the same retrieval analysis data or which can be regarded as the same is classified into one category to obtain each optimized classification; regarded as the same means that the retrieval conditions and the retrieval recommendation results can be regarded as the same in the actual retrieval process, such as the same meaning of different expressions of the retrieval conditions, the difference in the order of each retrieval recommendation result, etc.; in one embodiment, a retrieval equivalence judgment model can be established, and the expression of the retrieval equivalence judgment model is Where: q is the input data, which are two search analysis data to be compared; the output data is the search equivalence value YP(q); if the search equivalence requirements are met, they are considered to be equivalent or identical; the search equivalence judgment model is used to analyze the search analysis data to obtain the search equivalence values ​​between the search analysis data; and the search behavior data are classified according to the search equivalence values;

[0099] Evaluate each optimized classification to determine whether it meets the search criteria; mark the optimized classification that does not meet the search criteria as the target optimized classification;

[0100] Search optimization is performed based on target optimization classification.

[0101] That is, once it is clear in which aspects the current retrieval technology does not meet the retrieval requirements, the applied retrieval technology can be optimized through manual or other existing methods.

[0102] Methods for evaluating each optimization category include:

[0103] Identify and optimize the search behavior data corresponding to the classification, and obtain the corresponding search duration according to each search behavior data. The search duration is the duration for the user to find the required learning materials or the duration of the search behavior data. If the required learning data is not found, the search duration does not need to be considered, that is, it does not need to be identified.

[0104] Establish a behavior evaluation model. The expression of the behavior evaluation model is Where: pj is the input data, the input data is the search behavior data, j = 1, 2, ..., m, m is a positive integer; satisfying the search requirement means that the user has retrieved the required learning materials; otherwise, the search requirement is not met; T is the search time; T0 is the preset time, which is the time period within which the platform expects to complete the search; such as 1 minute, 2 minutes, 3 minutes, etc.; the output data is the behavior evaluation value XQ(pj);

[0105] Analyze each retrieval behavior data in the optimized classification through the behavior evaluation model to obtain the corresponding behavior evaluation value;

[0106] According to the formula Calculate the corresponding optimization evaluation value;

[0107] Where: PC is the optimized evaluation value;

[0108] The optimization classification whose optimization evaluation value is lower than the threshold X2 is marked as the target optimization classification.

[0109] A learning material retrieval evaluation autonomous optimization method, the method comprising:

[0110] Summarize various learning needs and the demand proportions corresponding to each learning need, and simulate and set up simulated retrieval sets according to each learning need;

[0111] Perform search simulation based on the simulated search set to obtain each simulated search result; analyze each simulated search result to determine each target resource channel; set the target resource channel account according to each target resource channel;

[0112] Connecting with various search channels; receiving the user's search conditions, searching the search conditions through various search channels, obtaining various initial search information, initially screening various initial search information, obtaining various candidate search information; evaluating various candidate search information, obtaining various search recommendation results, and displaying various search recommendation results to the user in columns according to whether they are charged;

[0113] Classify each search behavior data according to each search analysis data to obtain each optimized classification; the search analysis data includes search conditions and each search recommendation result;

[0114] Evaluate each optimization classification to determine whether it meets the search criteria; mark the optimization classifications that do not meet the search criteria as target optimization classifications; and perform search optimization based on the target optimization classifications.

[0115] The above formulas are all calculated by removing dimensions and taking numerical values. The formula is a formula that is closest to the actual situation obtained by collecting a large amount of data and performing software simulation. The preset parameters and preset thresholds in the formula are set by technical personnel in this field according to actual conditions or obtained by simulating a large amount of data.

[0116] The above embodiments are only used to illustrate the technical method of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical method of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical method of the present invention.

Claims

1. A learning material retrieval and evaluation autonomous optimization system, characterized in that: Includes account management module, search module and evaluation optimization module; The account management module is used to manage channel accounts, summarize various learning needs and the demand proportions corresponding to each learning need, and simulate and set a simulated search set according to each learning need; Perform search simulation based on the simulated search set to obtain each simulated search result; analyze each simulated search result to determine each target resource channel; set a target resource channel account according to each target resource channel; The retrieval module is used by users to retrieve learning materials, and includes an input unit and a retrieval unit; The input unit is used for the user to input the search condition and send the search condition to the search unit; The search unit is used to search for learning materials according to the search information; connect to various search channels; receive the user's search conditions, search the search conditions through various search channels, obtain various initial search information, perform initial screening on each of the initial search information, obtain various candidate search information; evaluate each candidate search information, obtain various search recommendation results, and display each search recommendation result to the user in columns according to whether it is charged; The evaluation and optimization module is used to perform search optimization based on the search behavior data of each user, obtain the search behavior data of each user and the corresponding search analysis data, and the search analysis data includes the search conditions and the search recommendation results; Classify each search behavior data according to each search analysis data to obtain each optimized classification; Evaluate each optimization classification to determine whether it meets the search criteria; mark the optimization classifications that do not meet the search criteria as target optimization classifications; and perform search optimization based on the target optimization classifications.

2. A learning material retrieval evaluation autonomous optimization system according to claim 1, characterized in that: Dynamically update the target resource channel account.

3. A learning material retrieval evaluation autonomous optimization system according to claim 1, characterized in that: The method for analyzing each simulated search result includes: Obtain each search recommendation result corresponding to each simulated search data; identify the source channel and source channel information corresponding to each search recommendation result; Classify each search recommendation result according to each source channel to obtain channel classification data corresponding to each source channel; and mark the corresponding demand proportion for each search recommendation result in each channel classification data; Identify the union part between the classified data of each channel, and identify the independent part in the classified data of each channel according to each of the union parts; mark each of the source channels with the independent part as the target resource channel; The remaining source channels are screened according to the respective union parts to obtain the respective target resource channels.

4. A learning material retrieval evaluation autonomous optimization system according to claim 3, characterized in that: Methods for filtering the remaining source channels according to the combined parts include: Mark each search recommendation result corresponding to the source channel as i, where i=1, 2, ..., n, and n is a positive integer; perform access evaluation on each search recommendation result according to the source channel information to obtain a corresponding access value; mark the access value as FEi; According to the formula Calculate the corresponding screening value; Where: PA is the screening value; βi is the demand proportion corresponding to the corresponding search recommendation result; The source channels are screened according to the screening values ​​to obtain the target resource channels.

5. The learning material retrieval evaluation autonomous optimization system according to claim 1, characterized in that: Methods for initially screening each initial search information include: An invalid evaluation model is established, and the expression of the invalid evaluation model is: Where: s is the input data, the input data is the search information and search conditions; the output data is the invalid evaluation value WP(s); Analyze each of the initial search information through an invalid evaluation model to obtain an invalid evaluation value corresponding to each of the initial search information; and remove each of the initial search information with an invalid evaluation value of 0; The remaining initial search information is identified, and the same initial search information is classified into one category to obtain each search information category; the initial search information in each search information category is evaluated and screened to obtain a candidate search information corresponding to each search information category.

6. A learning material retrieval evaluation autonomous optimization system according to claim 5, characterized in that: The method for evaluating and screening each initial search information in the search information classification includes: Estimate the access value corresponding to each initial search information; identify the target resource channel corresponding to each initial search information; evaluate each target resource channel according to the preset additional evaluation items to obtain the corresponding additional evaluation value; Calculate the corresponding screening value according to the formula PB = b1 × FZ-b2 × FE; Where: PB is the screening value; FZ is the additional value; FE is the access value; b1 and b2 are both proportional coefficients, with a value range of 0≤b1<1, 0 <b2≤1; The initial search information with the largest screening value is selected as the search information to be selected.

7. The learning material retrieval evaluation autonomous optimization system according to claim 1, characterized in that: The method of classifying each search behavior data according to each search analysis data includes: Establish a retrieval equivalence judgment model, the expression of the retrieval equivalence judgment model is: In the formula: q is the input data, the input data is the two search analysis data to be compared; the output data is the search equivalent value YP(q); The retrieval analysis data are analyzed by the retrieval equivalence judgment model to obtain the retrieval equivalence values ​​between the retrieval analysis data; the retrieval behavior data are classified according to the retrieval equivalence values ​​to obtain the optimized classifications.

8. The learning material retrieval evaluation autonomous optimization system according to claim 1, characterized in that: Methods for evaluating each optimization category include: Identify each search behavior data corresponding to the optimization classification, and obtain the corresponding search duration according to each search behavior data; A behavior evaluation model is established. The expression of the behavior evaluation model is: Where: pj is the input data, the input data is the retrieval behavior data, j = 1, 2, ..., m, m is a positive integer; T is the retrieval time; T0 is the preset time; the output data is the behavior evaluation value XQ (pj); Analyze each retrieval behavior data in the optimized classification through the behavior evaluation model to obtain the corresponding behavior evaluation value; According to the formula Calculate the corresponding optimization evaluation value; Where: PC is the optimized evaluation value; The optimization classification whose optimization evaluation value is lower than the threshold X2 is marked as the target optimization classification.

9. A method for autonomous optimization of learning material retrieval and evaluation, characterized in that: A learning material retrieval evaluation autonomous optimization system as claimed in any one of claims 1 to 8, the method comprising: Summarize various learning needs and the demand proportions corresponding to each learning need, and simulate and set up simulated retrieval sets according to each learning need; Perform search simulation based on the simulated search set to obtain each simulated search result; analyze each simulated search result to determine each target resource channel; set the target resource channel account according to each target resource channel; Connecting with various search channels; receiving the user's search conditions, searching the search conditions through various search channels, obtaining various initial search information, initially screening various initial search information, obtaining various candidate search information; evaluating various candidate search information, obtaining various search recommendation results, and displaying various search recommendation results to the user in columns according to whether they are charged; Classify each search behavior data according to each search analysis data to obtain each optimized classification; the search analysis data includes search conditions and each search recommendation result; Evaluate each optimization classification to determine whether it meets the search criteria; mark the optimization classifications that do not meet the search criteria as target optimization classifications; and perform search optimization based on the target optimization classifications.

Images