Autonomous iteration-based search feedback information determination method and device, equipment and medium

Abstract

This invention discloses a method, apparatus, device, and medium for determining search feedback information based on autonomous iteration. The method includes: acquiring the original query terms input by a target user and determining the query intent corresponding to the original query terms; determining a search engine combination based on the query intent; acquiring candidate information sources corresponding to the query intent based on the search engine combination and determining the information quality score corresponding to the candidate information sources; determining the target information source from the candidate information sources based on the information quality score; calculating the information entropy, information similarity, and resource consumption of the target information source; and determining whether a stopping condition is met. If the target information source meets the query stopping condition, generating a query feedback result based on the target information source and feeding the query feedback result back to the target user. If the target information source does not meet the query stopping condition, treating the target information source as a historical information source and returning the search engine combination determined based on the query intent. This ensures the accuracy of the feedback results.

Description

Technical Field

[0001] This invention relates to the field of artificial intelligence technology, and in particular to a method, apparatus, device, and medium for determining search feedback information based on autonomous iteration. Background Technology

[0002] With the development of computer technology, a large amount of data exists on the Internet. Users search through this vast amount of data to obtain data that corresponds to their query intent.

[0003] However, existing indexing methods obtain information from internet sources of varying quality and typically only perform simple relevance checks, failing to effectively filter out low-quality, outdated, or unreliable information. As a result, the accuracy and authority of the final search results do not meet users' needs. Summary of the Invention

[0004] This invention provides a method, apparatus, device, and medium for determining search feedback information based on autonomous iteration. By performing concurrent queries according to the query intent corresponding to the query terms, scoring the quality of the obtained information, and determining the target information source based on the scoring results, the final query feedback result is formed, thereby ensuring the accuracy and authority of the query feedback result.

[0005] According to one aspect of the present invention, a method for determining search feedback information based on autonomous iteration is provided, comprising:

[0006] Obtain the original query terms input by the target user and determine the query intent corresponding to the original query terms;

[0007] Based on the query intent, a combination of search engines is determined; candidate information sources corresponding to the query intent are obtained according to the combination of search engines; and an information quality score corresponding to the candidate information sources is determined.

[0008] Based on the information quality score, a target information source is determined from the candidate information sources. The information entropy, information similarity, and resource consumption of the target information source are calculated, and it is determined whether the stopping condition is met.

[0009] If the target information source meets the query stop condition, a query feedback result is generated based on the target information source and fed back to the target user; if the target information source does not meet the query stop condition, the target information source is used as a historical information source, and a search engine combination determined based on the query intent is returned.

[0010] According to another aspect of the present invention, a search feedback information determination device based on autonomous iteration is provided, comprising:

[0011] The query intent determination module is used to obtain the original query terms input by the target user and determine the query intent corresponding to the original query terms.

[0012] The information quality evaluation module is used to determine the search engine combination based on the query intent, obtain candidate information sources corresponding to the query intent according to the search engine combination, and determine the information quality score corresponding to the candidate information sources.

[0013] The stopping judgment module is used to determine the target information source from the candidate information sources based on the information quality score, calculate the information entropy, information similarity and resource consumption of the target information source, and determine whether the stopping condition is met.

[0014] The query result feedback module is used to generate query feedback results based on the target information source when the target information source meets the query stop condition, and to feed back the query feedback results to the target user.

[0015] The iterative query module is used to treat the target information source as a historical information source when the target information source does not meet the query stopping condition, and return a search engine combination determined based on the query intent.

[0016] According to another aspect of the present invention, an electronic device is provided, the electronic device comprising:

[0017] At least one processor; and

[0018] A memory communicatively connected to the at least one processor; wherein,

[0019] The memory stores a computer program that can be executed by the at least one processor, which enables the at least one processor to perform the autonomous iterative search feedback information determination method according to any embodiment of the present invention.

[0020] According to another aspect of the present invention, a computer-readable storage medium is provided, the computer-readable storage medium storing computer instructions, the computer instructions being configured to cause a processor to execute and implement the self-iterative search feedback information determination method described in any embodiment of the present invention.

[0021] The technical solution of this invention involves obtaining the original query terms input by the target user and determining the query intent corresponding to the original query terms; determining a search engine combination based on the query intent; obtaining candidate information sources corresponding to the query intent based on the search engine combination and determining the information quality score corresponding to the candidate information sources; determining the target information source from the candidate information sources based on the information quality score; calculating the information entropy, information similarity, and resource consumption of the target information source; and determining whether a stopping condition is met. If the target information source meets the query stopping condition, a query feedback result is generated based on the target information source and fed back to the target user. If the target information source does not meet the query stopping condition, the target information source is used as a historical information source, and the search engine combination determined based on the query intent is returned. Based on the above technical solution, by performing concurrent queries according to the query intent corresponding to the query terms, scoring the quality of the obtained information, and determining the target information source based on the scoring results, a final query feedback result is formed, thereby ensuring the accuracy and authority of the query feedback result.

[0022] It should be understood that the description in this section is not intended to identify key or essential features of the embodiments of the present invention, nor is it intended to limit the scope of the invention. Other features of the invention will become readily apparent from the following description. Attached Figure Description

[0023] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0024] Figure 1 This is a flowchart illustrating a method for determining search feedback information based on autonomous iteration, provided in an embodiment of the present invention.

[0025] Figure 2 A flowchart illustrating a method for determining search feedback information based on autonomous iteration, provided in an embodiment of the present invention;

[0026] Figure 3 This is a structural block diagram of a search feedback information determination device based on autonomous iteration provided in an embodiment of the present invention;

[0027] Figure 4 This is a schematic diagram of the structure of the electronic device provided in an embodiment of the present invention. Detailed Implementation

[0028] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0029] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of the invention described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0030] It should be noted that the acquisition, storage, use, and processing of data in the technical solution of this application all comply with the relevant provisions of national laws and regulations.

[0031] Example 1

[0032] Figure 1 This is a flowchart illustrating a method for determining search feedback information based on autonomous iteration, provided by an embodiment of the present invention. This embodiment is applicable to situations where the original query terms input by the user are indexed and the query feedback results are determined. This method can be executed by a device for determining search feedback information based on autonomous iteration. This device can be implemented in hardware and / or software and can be configured in an electronic device, such as a server or terminal device. Figure 1 As shown, the method includes:

[0033] S110. Obtain the original query terms input by the target user and determine the query intent corresponding to the original query terms.

[0034] The target user can be a user performing a data query. The original query term can be understood as the query word entered through the front-end interactive interface; it can be a text query or voice input by the user. The query intent can be the query direction corresponding to the query term, indicating the target user's data indexing purpose.

[0035] Specifically, this can be achieved through a front-end interactive interface, such as a web search box, an app input bar, or a voice interaction module, guiding the target user to input raw query terms. These raw query terms can be obtained through various input methods, including text and voice. After obtaining the raw query terms, lexical and syntactic analysis are used to structurally decompose the query terms, extracting keywords, entities, and grammatical relationships. Furthermore, by combining a pre-trained language model and domain knowledge graph, the semantic connotation of the query terms is analyzed to identify potential intent categories, such as "academic research," "business analysis," "news events," and "technical research," thereby obtaining the query intent corresponding to the target user.

[0036] S120. Determine a search engine combination based on the query intent, obtain candidate information sources corresponding to the query intent according to the search engine combination, and determine the information quality score corresponding to the candidate information sources.

[0037] In this invention, the search engine combination can be understood as a collection of multiple search engines corresponding to the query intent. Candidate information sources can be web pages corresponding to the original query terms obtained through parallel queries using the search engine combination. Information quality scoring can be understood as a score used to evaluate the quality of information in the information sources. It should be noted that the purpose of this invention is to construct an intelligent agent that precisely controls the entire process from information acquisition and filtering to integration, capable of simulating the research behavior of domain experts and automatically generating research reports efficiently and with high quality. The technical solution of this invention, based on an intelligent stopping mechanism of information entropy, effectively avoids invalid iterations. Under the same quality requirements, the average number of searches and LLM calls in this system is reduced by 40-50%, significantly reducing time and economic costs, and realizing a shift from "passive execution" to "active decision-making." Adaptive search, dynamic filtering, and intelligent convergence endow it with preliminary self-learning and self-evolution capabilities, representing the development direction of the next generation of AI agents.

[0038] Specifically, based on the type of query intent, such as academic research or news information, and the characteristics of the relevant field, a combination of search engines is selected and configured. It should be noted that this combination can include general search engines, vertical search engines, and database retrieval. Then, using customized query statements, searches are performed in parallel across the selected search engines to identify candidate information sources highly relevant to the query intent. To ensure information quality, a multi-dimensional evaluation model is used to comprehensively score the candidate information sources, including but not limited to the authority of the information source, the freshness of the content, the accuracy of the data, and user feedback. Through this technical solution, the information quality score of each candidate information source can be accurately determined, providing users with high-quality and reliable search results.

[0039] Based on the above technical solution, the step of determining the search engine combination based on the query intent includes: matching the query intent in a search engine knowledge base to determine a list of search engines corresponding to the query intent; determining the evaluation score corresponding to each search engine in the list of search engines; and filtering the search engine list based on the evaluation score to determine the search engine combination corresponding to the query intent.

[0040] The search intent includes at least one of the following: academic research, business analysis, news events, and technical research. The search engine knowledge base can be a pre-established database storing performance data corresponding to different search engines. This performance data can include data such as result relevance and authority ratings corresponding to the search engines. The search engine list can be understood as a list consisting of at least two search engines corresponding to the search intent. The evaluation score can be the current search engine's evaluation score in the domain of the query intent; for example, for a search intent related to academic research, the evaluation score could be an evaluation score corresponding to authority.

[0041] Specifically, matching can be performed in a pre-built search engine knowledge base based on the semantic features, domain attributes, and user preferences of the query intent. It should be noted that the search engine knowledge base is a pre-built database containing detailed information on various search engines, including their coverage, professional fields, and data update frequency. Then, based on the search intent, a list of search engines corresponding to the query intent is determined from the search engine knowledge base. Each search engine in the list is comprehensively evaluated to determine a corresponding evaluation score. This evaluation score can be determined based on multiple preset indicators, which may include dimensions such as retrieval accuracy, response speed, data source authority, and user satisfaction. Different query intents can correspond to different evaluation indicators. After calculating the evaluation scores corresponding to the indicators, a weighted scoring mechanism is used to determine the evaluation score corresponding to each search engine in the search engine list. Finally, the search engines are sorted from highest to lowest evaluation score, and the top K search engines are selected as the search engine combination.

[0042] The described technical solution maximizes the fulfillment of query intent while balancing retrieval efficiency and information quality, providing users with an efficient and accurate search experience. For example, a database is maintained to record the historical performance of various search engines, such as Google Scholar, Bing, DuckDuckGo, and Perplexity, under different query types, including result relevance and authority scores. Then, based on the current query type, the top-K search engines with the best historical performance are selected from the knowledge base for concurrent searches. It should be noted that for comprehensive queries, a weighted fusion strategy can be used. This involves separately determining the search engine combinations corresponding to the query intent, weighting and summing their corresponding evaluation scores, and then selecting the top-K search engines with the highest overall scores as the final search engine combination.

[0043] S130. Determine the target information source from the candidate information sources based on the information quality score, calculate the information entropy, information similarity and resource consumption of the target information source, and determine whether the stopping condition is met.

[0044] The target information source can be any information source that meets the filtering criteria obtained from candidate information sources. The query stopping condition can be understood as a pre-set condition used to stop the query.

[0045] Specifically, based on the information quality score, a sorting algorithm can be used to rank candidate information sources in descending order, prioritizing those with higher scores as target information sources. Alternatively, a scoring threshold can be set; information sources with scores reaching or exceeding this threshold are considered target information sources. The system then determines whether the identified target information sources meet the query stopping conditions. These conditions include information completeness (measuring whether the target information source covers all the key points required for the query), information novelty (measuring whether the target information source contains the most up-to-date data), and specific user-defined requirements, such as source authority and length. If the target information source meets all preset conditions, the query process stops, and the final result is output; otherwise, the search scope continues to be filtered or expanded until a target information source that meets the conditions is found.

[0046] Based on the above technical solution, the step of determining the target information source from the candidate information sources according to the information quality score includes: obtaining a quality score threshold corresponding to the query intent; comparing the information quality score of each information source in the candidate information sources with the quality score threshold; and selecting the information source whose information quality score is greater than the quality score threshold as the target information source.

[0047] The quality scoring threshold can be a pre-set quality score value used to determine the target information source.

[0048] Specifically, the quality score threshold corresponding to the query intent is obtained. It should be noted that the quality score threshold can be determined based on statistical analysis of historical query data, for example, extracting the generally accepted information quality level for similar query intents; or it can be based on domain expert knowledge, pre-setting a minimum quality standard for a specific query intent. After obtaining the quality score threshold, the candidate information source list is automatically traversed, and the information quality score of each information source is compared with the threshold. Information sources with information quality scores greater than the preset quality score threshold are filtered out as target information sources. The technical solution of this embodiment effectively eliminates low-quality or irrelevant information, ensuring the accuracy and reliability of the search results and providing users with a high-quality information service experience.

[0049] Based on the above technical solution, the calculation of information entropy, information similarity, and resource consumption with the target information source, and the determination of whether the stopping condition is met, includes: calculating the conditional information entropy increment and information similarity based on the historical information source and the target information source; and determining whether the query stopping condition is met based on the conditional information entropy increment and the information similarity.

[0050] Specifically, the conditional information entropy increment is first calculated based on the distribution patterns of data items in historical information sources and the expected characteristics of the target information source. The conditional information entropy increment reflects the degree of uncertainty change after the introduction of the target information source. A significant increment indicates that the target information source carries a large amount of new information. Information similarity is quantified by comparing the similarity between historical and target information sources in terms of data structure and feature dimensions; higher similarity indicates a stronger correlation. The conditional information entropy increment and information similarity are then comprehensively evaluated based on a preset threshold standard. If the conditional information entropy increment is below the threshold and the information similarity is above the threshold, it is determined that the current query has sufficiently mined effective information, meeting the query termination condition, and the query process can be terminated.

[0051] Based on the above technical solution, the step of determining whether the query stop condition is met according to the historical information source and the target information source includes: determining whether the conditional information entropy increment is lower than a preset entropy threshold, and determining whether the information similarity is greater than a preset similarity threshold; if the conditional information entropy increment is lower than the preset entropy threshold and the information similarity is greater than the preset similarity threshold, then determining that the target information source meets the query stop condition.

[0052] Information entropy is used to characterize how much "unexpectedness" or "reduction in uncertainty" is brought about by new information. Information similarity is used to determine the degree of similarity between the target information source and historical information sources.

[0053] Specifically, by calculating the information entropy between historical and target information sources, the differences and uncertainties in the information they carry are quantified. If the information entropy gradually converges to a stable value, it indicates that the newly acquired target information source has not brought significant new information, i.e., the information increment approaches zero. At this point, the target information source can be determined to meet the query stopping condition, avoiding invalid retrieval. Simultaneously, the information similarity between the two sources is calculated, using algorithms such as cosine similarity and Jaccard similarity coefficient to measure the degree of overlap in information content. When the information similarity exceeds a preset similarity threshold, it indicates that the target information source and historical information source are highly repetitive, and further retrieval is unlikely to obtain more valuable information. Based on the above two judgment methods, it is jointly determined that the target information source meets the query stopping condition. For example, all extracted and fused contextual information is considered as an information set S_old. For the high-quality information S_new acquired in a new iteration, the conditional information entropy increment it brings is calculated: H(S_new|S_old). This value reflects how much "unexpected" or "reduction in uncertainty" the new information brings. A convergence threshold $\epsilon$ is set. When $H(S_{new} | S_{old}) \< \\epsilon$, it indicates that the new round of search has not brought significant information entropy and has approached an "information saturation" state. A multi-objective decision model is constructed by combining the information entropy increment, the similarity between the newly generated query and the historical query, and the total resource consumption. A stopping signal is triggered when any two of these conditions are met (e.g., "information entropy is below a threshold" and "the new query is highly similar to the old query").

[0054] It should be noted that information entropy, in information theory, is used to measure the uncertainty of a random variable. Here, it measures the novelty and information content of a set of information. Conditional information entropy $H(Y|X)$ represents the uncertainty of Y given that $X$.

[0055] S140. If the target information source meets the query stop condition, generate a query feedback result based on the target information source and send the query feedback result back to the target user.

[0056] S150. If the target information source does not meet the query stop condition, the target information source is used as a historical information source, and the search engine combination determined based on the query intent is returned.

[0057] The query feedback results can be information displayed to the user, including links to pages from various information sources within the target information source, or structured text obtained after structuring the target information source. Historical information sources can be those that do not meet the query stopping criteria.

[0058] Specifically, the target information source is analyzed and integrated to extract core content elements, such as key data, conclusions, and solutions, and then transformed into structured feedback text using natural language generation technology. For example, for academic queries, the core viewpoints of multiple highly authoritative papers are summarized to generate a review feedback that includes research background, methods, and results; for product queries, information such as price, parameters, and user reviews is integrated to generate a comparative analysis report. Based on user history and preferences, the feedback results are personalized and optimized, such as adjusting the use of technical terminology or adding visualization charts. Finally, the query feedback results can be fed back to the target user's terminal in real time through preset communication protocols, such as HTTP API, email push, and message queues, and user interaction data, such as click-through rate and dwell time, is recorded to optimize subsequent query strategies. If the current target information source is determined not to meet the preset query stop conditions, it means that the information source has not fully covered the query needs. In this case, the current target information source is dynamically included in the historical information source set as a reference benchmark for subsequent queries. Based on the user's initial query intent, combined with historical query logs and semantic analysis technology, the search engine combination strategy is rematched, and the updated search engine combination is returned to execute a new round of queries, forming a closed-loop optimization process of "evaluation-iteration-query".

[0059] The technical solution of this invention involves acquiring the original query terms input by the target user, determining the query intent corresponding to the original query terms, determining a search engine combination based on the query intent, acquiring candidate information sources corresponding to the query intent based on the search engine combination, determining the information quality score corresponding to the candidate information sources, and then determining the target information source from the candidate information sources based on the information quality score. The solution also calculates the information entropy, information similarity, and resource consumption of the target information source, determines whether a stopping condition is met, and if the target information source meets the query stopping condition, generates a query feedback result based on the target information source and sends the query feedback result back to the target user. Based on this technical solution, by performing concurrent queries according to the query intent corresponding to the query terms, scoring the quality of the acquired information, and determining the target information source based on the score results, a final query feedback result is formed, thereby ensuring the accuracy and authority of the query feedback result.

[0060] Example 2

[0061] Figure 2 This is a flowchart illustrating a method for determining search feedback information based on autonomous iteration, provided by an embodiment of the present invention. This embodiment further explains the technical solution for determining the information quality score corresponding to the candidate information source, based on the above technical solution. Figure 2 As shown, the method includes:

[0062] S210. For each of the candidate information sources, determine the information evaluation score corresponding to each evaluation dimension based on the information evaluation dimension.

[0063] The information evaluation dimensions can be multiple pre-set evaluation dimensions used to score information sources. These dimensions must include at least relevance, authority, timeliness, and completeness. The information evaluation score can be the corresponding score for each evaluation dimension.

[0064] Specifically, in terms of relevance, natural language processing techniques are used to analyze the semantic matching degree between the information source content and the query intent. A relevance score of 0-10 is given through methods such as keyword co-occurrence and semantic similarity calculation. The authority dimension examines indicators such as the publishing institution, author qualifications, and citation count of the information source, combined with a domain knowledge base to determine and score authority. The timeliness dimension scores based on the interval between the information's publication or update time and the current time; the shorter the interval, the higher the score. The completeness dimension assesses the information source's coverage of the query topic, whether it includes key elements and detailed descriptions, thereby determining the completeness score. Finally, the scores from all dimensions are combined to obtain a comprehensive information evaluation score for each information source.

[0065] S220. Determine the weight value corresponding to the evaluation dimension based on the query intent, and determine the information quality score corresponding to each information source based on the weight value and the information evaluation score corresponding to each evaluation dimension.

[0066] The weight value can be a weight coefficient corresponding to each evaluation dimension, and different weight coefficients can be set for the evaluation dimensions according to different query intentions.

[0067] Specifically, based on the semantic features and domain attributes of the query intent, the analytic hierarchy process (AHP) is used to determine the weight values ​​for each evaluation dimension. For example, academic queries may place greater emphasis on authority (weight 0.4) and completeness (weight 0.3), while news queries focus on timeliness (weight 0.5) and relevance (weight 0.3), with the authority weight reduced to 0.2. The scores (0-10 points) of each information source in the dimensions of relevance, authority, timeliness, and completeness are multiplied by their corresponding weights and then summed. For example, if an information source's scores in the four dimensions are 8, 7, 9, and 6, and the weights are 0.3, 0.2, 0.3, and 0.2 respectively, then its overall score is 8×0.3 + 7×0.2 + 9×0.3 + 6×0.2 = 7.7 points. Finally, the weighted total score for each information source is output as the information quality score.

[0068] Based on the above technical solution, the step of determining the information evaluation score corresponding to each evaluation dimension of the information source includes: extracting information text data corresponding to the information source, and determining the information evaluation score corresponding to the relevance dimension based on the information text data and the original query terms; determining the webpage domain name and webpage referencing relationship corresponding to the information source, and determining the information evaluation score corresponding to the authority dimension based on the webpage domain name and webpage referencing relationship; obtaining the page update time corresponding to the information source, and calculating the text repetition between the information sources, and determining the information evaluation score corresponding to the timeliness dimension based on the page update time and the text repetition; determining the text structure information corresponding to the information source, and determining the information evaluation score corresponding to the completeness dimension based on the text result information.

[0069] The information text data can be the text data in the webpage of the information source. The webpage domain name can be understood as the webpage suffix of the information source, used to identify the page type of the information source. Webpage referencing relationships can be the relationships in which information sources are referenced. Page update time can be understood as the last update time of the page in the information source. Text repetition can be the degree of text repetition between different information sources. Text structure information can be understood as the structured information of the text in the information source, which can include paragraphs, headings, data, etc.

[0070] Specifically, regarding the relevance dimension, the core text data of the information source is extracted, and the semantic matching degree between the text and the original query term is calculated using the TF-IDF algorithm or the BERT semantic model. For example, if the query term is "artificial intelligence application," the frequency of keyword occurrence, semantic similarity, and contextual relevance in the text will be analyzed to generate a relevance score of 0-10. Text with high-frequency matching and semantic consistency scores higher, while text containing only scattered keywords scores lower.

[0071] Regarding the authority dimension, the authority of information sources is assessed by analyzing their webpage domains, such as .gov and .edu suffixes, and external citations, using link analysis algorithms like PageRank. For example, information sources cited by multiple authoritative websites, or domains from government or academic institutions, will receive higher authority scores; while personal blogs or low-quality forums will score lower.

[0072] Regarding timeliness, by capturing the page update time of information sources and combining it with time markers in the text content, such as news release dates, recently updated content is prioritized. Duplicate or outdated information is identified by calculating the text repetition between information sources, such as the Jaccard similarity coefficient. For example, information sources updated within the last 3 months and with a repetition rate below 30% receive higher scores, while those not updated for more than a year or with a repetition rate exceeding 70% receive lower scores.

[0073] Regarding the completeness dimension, the text structure information of the information source is analyzed, including paragraph hierarchy, heading completeness, and coverage of key elements. For example, texts with a clear structure that includes an introduction, body, and conclusion score higher, while those containing only fragmented content score lower. Through pre-defined completeness template matching, the degree to which the text covers the query topic can be quantitatively evaluated.

[0074] For example, a four-dimensional evaluation model is designed to score each information source: Content Relevance: The semantic matching degree between the page content and the query is calculated using a vector similarity algorithm. Source Authority: Evaluation is based on preset domain weights (e.g., .edu, .gov have high weights) and page citation relationships (e.g., PageRank). Timeliness & Novelty: Scoring is based on the page's publication / update time and the degree of repetition of content with existing information, using a time decay function to reduce the weight of older information. Information Completeness: The completeness of the content is initially judged by analyzing the text structure (e.g., whether there are clear paragraphs, headings, and data). Dynamic Threshold Filtering: The comprehensive score uses a weighted summation formula: Score = α × R + β × A + γ × T + δ × C, where the weights α, β, γ, and δ can be dynamically adjusted according to the query type (e.g., academic queries place more emphasis on authority β, while news queries place more emphasis on timeliness γ). Only information sources with a Score higher than the dynamically adjusted threshold are accepted.

[0075] The technical solution of this invention involves obtaining the original query terms input by the target user, determining the query intent corresponding to the original query terms, determining a search engine combination based on the query intent, obtaining candidate information sources corresponding to the query intent based on the search engine combination, determining the information quality score corresponding to the candidate information sources, and then determining the target information source from the candidate information sources based on the information quality score. The solution also calculates the information entropy, information similarity, and resource consumption of the target information source, determines whether a stopping condition is met, and if the target information source meets the query stopping condition, generates a query feedback result based on the target information source and sends the query feedback result back to the target user. Based on this technical solution, by performing concurrent queries based on the query intent corresponding to the query terms, scoring the quality of the obtained information, and determining the target information source based on the score results, a final query feedback result is formed, thereby ensuring the accuracy and authority of the query feedback result.

[0076] Example 3

[0077] Figure 3This is a structural block diagram of a search feedback information determination device based on autonomous iteration, provided as an embodiment of the present invention. Figure 3 As shown, the device includes: a query intent determination module 310, an information quality evaluation module 320, a stop judgment module 330, a query result feedback module 340, and an iterative query module 350; wherein.

[0078] The query intent determination module 310 is used to obtain the original query terms input by the target user and determine the query intent corresponding to the original query terms;

[0079] The information quality evaluation module 320 is used to determine the search engine combination based on the query intent, obtain candidate information sources corresponding to the query intent according to the search engine combination, and determine the information quality score corresponding to the candidate information sources.

[0080] The stopping judgment module 330 is used to determine the target information source from the candidate information sources based on the information quality score, calculate the information entropy, information similarity and resource consumption of the target information source, and determine whether the stopping condition is met.

[0081] The query result feedback module 340 is used to generate a query feedback result based on the target information source when the target information source meets the query stop condition, and to feed the query feedback result back to the target user.

[0082] The iterative query module 350 is used to treat the target information source as a historical information source and return a search engine combination determined based on the query intent when the target information source does not meet the query stopping condition.

[0083] Based on the above technical solution, the information quality evaluation module is used to match the query intent in the search engine knowledge base, determine the list of search engines corresponding to the query intent, wherein the search intent includes at least one of the following: academic research, business analysis, news events, and technical research; determine the evaluation score corresponding to each search engine in the list of search engines, and filter the search engines from the list of search engines according to the evaluation score to determine the combination of search engines corresponding to the query intent.

[0084] Based on the above technical solution, the information quality evaluation module is used to determine the information evaluation score corresponding to each information source among the candidate information sources according to the information evaluation dimensions, wherein the information evaluation dimensions include at least the relevance dimension, authority dimension, timeliness dimension, and completeness dimension; determine the weight value corresponding to the evaluation dimension according to the query intent; and determine the information quality score corresponding to each information source according to the weight value and the information evaluation score corresponding to each evaluation dimension.

[0085] Based on the above technical solution, the information quality evaluation module is used to extract information text data corresponding to the information source, determine the information evaluation score corresponding to the relevance dimension based on the information text data and the original query terms; determine the webpage domain name and webpage referencing relationship corresponding to the information source, and determine the information evaluation score corresponding to the authority dimension based on the webpage domain name and webpage referencing relationship; obtain the page update time corresponding to the information source, calculate the text repetition between the information sources, and determine the information evaluation score corresponding to the timeliness dimension based on the page update time and the text repetition; determine the text structure information corresponding to the information source, and determine the information evaluation score corresponding to the completeness dimension based on the text result information.

[0086] Based on the above technical solution, the stop judgment module is used to obtain a quality score threshold corresponding to the query intent, compare the information quality score of each information source in the candidate information sources with the quality score threshold, and take the information source with the information quality score greater than the quality score threshold as the target information source.

[0087] Based on the above technical solution, the stop judgment module is used to calculate the conditional information entropy increment and information similarity according to the historical information source and the target information source; and to determine whether the query stop condition is met based on the conditional information entropy increment and the information similarity.

[0088] Based on the above technical solution, the stop judgment module is used to determine whether the conditional information entropy increment is lower than a preset entropy threshold, and to determine whether the information similarity is greater than a preset similarity threshold; if the conditional information entropy increment is lower than the preset entropy threshold and the information similarity is greater than the preset similarity threshold, the target information source is determined to meet the query stop condition.

[0089] The technical solution of this invention involves obtaining the original query terms input by the target user, determining the query intent corresponding to the original query terms, determining a search engine combination based on the query intent, obtaining candidate information sources corresponding to the query intent based on the search engine combination, determining the information quality score corresponding to the candidate information sources, and then determining the target information source from the candidate information sources based on the information quality score. The solution also calculates the information entropy, information similarity, and resource consumption of the target information source, determines whether a stopping condition is met, and if the target information source meets the query stopping condition, generates a query feedback result based on the target information source and sends the query feedback result back to the target user. Based on this technical solution, by performing concurrent queries based on the query intent corresponding to the query terms, scoring the quality of the obtained information, and determining the target information source based on the score results, a final query feedback result is formed, thereby ensuring the accuracy and authority of the query feedback result.

[0090] The search feedback information determination device based on autonomous iteration provided in the embodiments of the present invention can execute the search feedback information determination method based on autonomous iteration provided in any embodiment of the present invention, and has the corresponding functional modules and beneficial effects of the execution method.

[0091] Example 4

[0092] Figure 4 A schematic diagram of an electronic device 10, which can be used to implement embodiments of the present invention, is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital processors, cellular phones, smartphones, wearable devices (e.g., helmets, glasses, watches, etc.), and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely illustrative and are not intended to limit the implementation of the invention described and / or claimed herein.

[0093] like Figure 4As shown, the electronic device 10 includes at least one processor 11 and a memory, such as a read-only memory (ROM) 12 or a random access memory (RAM) 13, communicatively connected to the at least one processor 11. The memory stores computer programs executable by the at least one processor. The processor 11 can perform various appropriate actions and processes based on the computer program stored in the ROM 12 or loaded from storage unit 18 into the RAM 13. The RAM 13 can also store various programs and data required for the operation of the electronic device 10. The processor 11, ROM 12, and RAM 13 are interconnected via a bus 14. An input / output (I / O) interface 15 is also connected to the bus 14.

[0094] Multiple components in electronic device 10 are connected to I / O interface 15, including: input unit 16, such as keyboard, mouse, etc.; output unit 17, such as various types of displays, speakers, etc.; storage unit 18, such as disk, optical disk, etc.; and communication unit 19, such as network card, modem, wireless transceiver, etc. Communication unit 19 allows electronic device 10 to exchange information / data with other devices through computer networks such as the Internet and / or various telecommunications networks.

[0095] Processor 11 can be a variety of general-purpose and / or special-purpose processing components with processing and computing capabilities. Some examples of processor 11 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various special-purpose artificial intelligence (AI) computing chips, various processors running machine learning model algorithms, digital signal processors (DSPs), and any suitable processor, controller, microcontroller, etc. Processor 11 performs the various methods and processes described above, such as the method of determining based on autonomous iterative search feedback information.

[0096] In some embodiments, the autonomous iterative search feedback information determination method can be implemented as a computer program tangibly contained in a computer-readable storage medium, such as storage unit 18. In some embodiments, part or all of the computer program can be loaded and / or installed on electronic device 10 via ROM 12 and / or communication unit 19. When the computer program is loaded into RAM 13 and executed by processor 11, one or more steps of the autonomous iterative search feedback information determination method described above can be performed. Alternatively, in other embodiments, processor 11 can be configured to perform the autonomous iterative search feedback information determination method by any other suitable means (e.g., by means of firmware).

[0097] Various embodiments of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field-programmable gate arrays (FPGAs), application-specific integrated circuits (ASICs), application-specific standard products (ASSPs), systems-on-a-chip (SoCs), payload-programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments may include implementations in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which may be a dedicated or general-purpose programmable processor, capable of receiving data and instructions from a storage system, at least one input device, and at least one output device, and transmitting data and instructions to the storage system, the at least one input device, and the at least one output device.

[0098] Computer programs used to implement the methods of the present invention may be written in any combination of one or more programming languages. These computer programs may be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, such that when executed by the processor, the computer programs cause the functions / operations specified in the flowcharts and / or block diagrams to be performed. The computer programs may be executed entirely on a machine, partially on a machine, or as a standalone software package, partially on a machine and partially on a remote machine, or entirely on a remote machine or server.

[0099] In the context of this invention, a computer-readable storage medium can be a tangible medium that may contain or store a computer program for use by or in conjunction with an instruction execution system, apparatus, or device. A computer-readable storage medium may include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination thereof. Alternatively, a computer-readable storage medium may be a machine-readable signal medium. More specific examples of machine-readable storage media include electrical connections based on one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fibers, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof.

[0100] To provide interaction with a user, the systems and techniques described herein can be implemented on an electronic device having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and pointing device (e.g., a mouse or trackball) through which the user provides input to the electronic device. Other types of devices can also be used to provide interaction with the user; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including sound input, voice input, or tactile input).

[0101] The systems and technologies described herein can be implemented in computing systems that include backend components (e.g., as data servers), or middleware components (e.g., application servers), or frontend components (e.g., user computers with graphical user interfaces or web browsers through which users can interact with implementations of the systems and technologies described herein), or any combination of such backend, middleware, or frontend components. The components of the system can be interconnected via digital data communication of any form or medium (e.g., communication networks). Examples of communication networks include local area networks (LANs), wide area networks (WANs), blockchain networks, and the Internet.

[0102] A computing system can include clients and servers. Clients and servers are generally located far apart and typically interact through communication networks. The client-server relationship is created by computer programs running on the respective computers and having a client-server relationship with each other. The server can be a cloud server, also known as a cloud computing server or cloud host, which is a hosting product within the cloud computing service system to address the shortcomings of traditional physical hosts and VPS services, such as high management difficulty and weak business scalability.

[0103] It should be understood that the various forms of processes shown above can be used, with steps reordered, added, or deleted. For example, the steps described in this invention can be executed in parallel, sequentially, or in different orders, as long as the desired result of the technical solution of this invention can be achieved, and this is not limited herein.

[0104] The specific embodiments described above do not constitute a limitation on the scope of protection of this invention. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this invention should be included within the scope of protection of this invention.

Claims

1. A method for determining search feedback information based on autonomous iteration, characterized in that, include: Obtain the original query terms input by the target user and determine the query intent corresponding to the original query terms; Based on the query intent, a combination of search engines is determined; candidate information sources corresponding to the query intent are obtained according to the combination of search engines; and an information quality score corresponding to the candidate information sources is determined. Based on the information quality score, a target information source is determined from the candidate information sources. The information entropy, information similarity, and resource consumption of the target information source are calculated, and it is determined whether the stopping condition is met. If the target information source meets the query stop condition, a query feedback result is generated based on the target information source, and the query feedback result is sent back to the target user. If the target information source does not meet the query stop condition, the target information source is treated as a historical information source, and the search engine combination determined based on the query intent is returned.

2. The method according to claim 1, characterized in that, The calculation of the information entropy, information similarity, and resource consumption of the target information source, and the determination of whether the stopping condition is met, includes: Calculate the conditional information entropy increment and information similarity based on the historical information source and the target information source; Based on the conditional information entropy increment and the information similarity, it is determined whether the query stopping condition is met.

3. The method according to claim 2, characterized in that, The step of determining whether the query stopping condition is met based on the conditional information entropy increment and the information similarity includes: Determine whether the conditional information entropy increment is lower than a preset entropy threshold, and determine whether the information similarity is greater than a preset similarity threshold; If the increment of the conditional information entropy is lower than the preset entropy threshold and the information similarity is greater than the preset similarity threshold, the target information source is determined to meet the query stop condition.

4. The method according to claim 1, characterized in that, The step of determining the search engine combination based on the query intent includes: The search intent is matched against the search engine knowledge base to determine a list of search engines corresponding to the search intent, wherein the search intent includes at least one of the following categories: academic research, business analysis, news events, and technical research. Determine the evaluation score corresponding to each search engine in the search engine list, and filter the search engine list based on the evaluation score to determine the combination of search engines corresponding to the query intent.

5. The method according to claim 1, characterized in that, Determining the information quality score corresponding to the candidate information source includes: For each of the candidate information sources, an information evaluation score is determined based on the information evaluation dimensions, which include at least the relevance dimension, authority dimension, timeliness dimension, and completeness dimension. Based on the query intent, a weight value corresponding to the evaluation dimension is determined, and based on the weight value and the information evaluation score corresponding to each evaluation dimension, an information quality score corresponding to each information source is determined.

6. The method according to claim 5, characterized in that, The step of determining the information source and the information evaluation score corresponding to each evaluation dimension based on the information evaluation dimensions includes: Extract the information text data corresponding to the information source, and determine the information evaluation score corresponding to the relevance dimension based on the information text data and the original query terms; Determine the webpage domain name and webpage reference relationship corresponding to the information source, and determine the information evaluation score corresponding to the authority dimension based on the webpage domain name and webpage reference relationship; Obtain the page update time corresponding to the information source, calculate the text repetition between the information sources, and determine the information evaluation score corresponding to the timeliness dimension based on the page update time and the text repetition. Determine the text structure information corresponding to the information source, and determine the information evaluation score corresponding to the integrity dimension based on the text result information.

7. The method according to claim 1, characterized in that, The step of determining the target information source from the candidate information sources based on the information quality score includes: Obtain the quality score threshold corresponding to the query intent, and compare the information quality score of each information source in the candidate information sources with the quality score threshold; Information sources whose information quality score is greater than the quality score threshold are designated as target information sources.

8. A search feedback information determination device based on autonomous iteration, characterized in that, include: The query intent determination module is used to obtain the original query terms input by the target user and determine the query intent corresponding to the original query terms. The information quality evaluation module is used to determine the search engine combination based on the query intent, obtain candidate information sources corresponding to the query intent according to the search engine combination, and determine the information quality score corresponding to the candidate information sources. The stopping judgment module is used to determine the target information source from the candidate information sources based on the information quality score, calculate the information entropy, information similarity and resource consumption of the target information source, and determine whether the stopping condition is met. The query result feedback module is used to generate query feedback results based on the target information source when the target information source meets the query stop condition, and to feed back the query feedback results to the target user. The iterative query module is used to treat the target information source as a historical information source when the target information source does not meet the query stopping condition, and return a search engine combination determined based on the query intent.

9. An electronic device, characterized in that, The electronic device includes: At least one processor; and A memory communicatively connected to the at least one processor; wherein, The memory stores a computer program that can be executed by the at least one processor, the computer program being executed by the at least one processor to enable the at least one processor to perform the autonomous iterative search feedback information determination method according to any one of claims 1-7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer instructions that cause a processor to execute the search feedback information determination method based on autonomous iteration as described in any one of claims 1-7.

Images