An information consultation strategy intelligent optimization system and method based on artificial intelligence

Abstract

This invention discloses an intelligent optimization system and method for information consulting strategies based on artificial intelligence, relating to the field of consulting strategy optimization technology. It establishes a corporate characteristic database, compares the characteristics of a new enterprise with those in the database to obtain the enterprise type, and considers the similarity between the new enterprise and the enterprises in the database when the new enterprise seeks information consulting, thus obtaining a preliminary recommendation level for search results obtained from the database. It also considers the relevance between historical search keywords and search results, comprehensively considering both the relevance and preliminary recommendation level to provide search results for the new enterprise. The search results are tagged, and historical satisfaction with the search results is considered. Weights are assigned to the relevance and preliminary recommendation level, and the database is updated in real time. This invention, by incorporating satisfaction feedback from similar enterprises regarding search results, makes the recommended search results more closely aligned with the actual business scenarios of the new enterprise.

Description

Technical Field

[0001] This invention relates to the field of consulting strategy optimization technology, specifically to an intelligent optimization system and method for information consulting strategies based on artificial intelligence. Background Technology

[0002] In the process of enterprise operation and development, information consulting is an important guarantee for assisting decision-making and supporting business expansion. Enterprises are increasingly reliant on precise information consulting services tailored to their specific needs. However, the current information consulting field generally suffers from a core technological bottleneck: insufficient accuracy in matching enterprise classification with consulting services, making it difficult to meet enterprises' core demand for efficiently obtaining suitable information. Traditional information consulting strategies lack scientific and systematic technical support in the enterprise classification stage and have not established a deep mining mechanism based on authorized historical enterprise data. Most methods rely on only a few superficial dimensions to classify enterprises, without standardizing enterprise data or extracting core enterprise characteristics and constructing a reasonable feature interval system. This results in superficial enterprise classification results that fail to accurately capture the deep connections between different enterprises in terms of business attributes, needs preferences, and development stages. This core technological problem directly leads to a chain of drawbacks: when a new enterprise initiates information consulting, the system cannot match effective experience from similar enterprises based on accurate enterprise classification and can only adopt a generalized information recommendation logic. As a result, the recommended search results are often out of touch with the actual needs of the enterprise, with a large amount of irrelevant information filling the consulting results. This not only consumes a lot of information screening time for enterprises and reduces consulting efficiency, but may also lead to enterprises obtaining incorrect or useless information, affecting the scientific nature of decision-making. Due to discrepancies in the basic classification of enterprises, subsequent information matching lacks a reliable basis, further exacerbating the inefficiency and inadequacy of consulting services. This severely restricts the supporting role of information consulting services in enterprise development, and there is an urgent need for targeted technical solutions to break through this core bottleneck. Summary of the Invention

[0003] The purpose of this invention is to provide an intelligent optimization system and method for information consulting strategies based on artificial intelligence, so as to solve the problems mentioned in the background art.

[0004] To address the aforementioned technical problems, this invention provides the following technical solution: an intelligent optimization method for information consulting strategies based on artificial intelligence, comprising the following steps: S1. Classify enterprises based on historical data and establish an enterprise characteristic database; S2. When receiving information inquiries from a new enterprise, compare the new enterprise's characteristics with those in the enterprise characteristic database to obtain the new enterprise's enterprise type; S3. Pre-set keyword database: When a new company seeks information, the similarity between the new company and companies in the database is considered to obtain the preliminary recommendation level of the search results obtained by the companies in the database. S4. Retrieve keywords from the new company's historical searches, consider the relevance between the keywords and search results, and comprehensively consider the relevance between the keywords and search results and the initial recommendation of the search results to provide search results for the new company. S5. Before recommending search results, mark the search results, consider the historical satisfaction of the search results, and set weights for relevance and initial recommendation. S6. Cumulative keyword search counts, with real-time database updates.

[0005] After filtering the search results, the historical search keywords of the new company are retrieved, and the relevance between the keywords and the search results is analyzed. During the analysis, the search records of similar companies are called up, and the number of times the target result is included in the search results when similar companies use the corresponding keywords is statistically analyzed. The relevance between a single keyword and its corresponding result is determined, and then the overall relevance level is obtained by combining the relevance of all relevant keywords.

[0006] Furthermore, in step S1, after authorization, enterprise data is acquired based on historical data. This data is preprocessed and normalized to obtain enterprise characteristics. These characteristics are then divided into characteristic intervals, and enterprises are categorized into X types, resulting in an enterprise characteristic database for each type. Through in-depth mining of historical enterprise data after authorization, invalid information is removed through standardization, core characteristics are extracted, and an enterprise characteristic database is constructed by scientifically dividing characteristic intervals, achieving systematic and precise enterprise classification. When consulting with new enterprises, classification is completed based on a precise comparison of their own characteristics with the classification standards in the database, ensuring that subsequent information recommendations strictly anchor to the effective experience and common needs of similar enterprises. This significantly reduces the interference of irrelevant information on the consultation results, making the information consultation direction more aligned with the actual business scenarios and core needs of enterprises. It helps enterprises quickly filter and obtain suitable key information, effectively improving the efficiency and accuracy of information consultation, and laying a solid foundation for subsequent intelligent recommendations throughout the entire process.

[0007] Furthermore, in step S2, upon receiving information inquiries from a new enterprise, the enterprise data of the new enterprise is obtained. After preprocessing and normalizing the enterprise data, the enterprise characteristics of the new enterprise are obtained. The enterprise characteristics of the new enterprise are {A1, A2, ..., A...}. n ,…,A N}, where N represents the number of firm features, A n This represents the nth enterprise characteristic of a new enterprise. The enterprise characteristic database is accessed, and the standard enterprise characteristic for the xth type of enterprise is {B1_ x B2_ x ,…,B n _ x ,…,B N _x}, where B n _ x B represents the nth standard firm characteristic of firm x. n Let C be the average of the nth firm characteristic of all firms in the x-th firm category, and then we can obtain the dissimilarity C between the nth characteristic of the new firm and the firms in the x-th firm category. n _ x C n_x =|(A n -B n _ x ) / B n _ x Substitute each of the following into n=1,2,…,N to obtain the dissimilarity of N features between the new firm and the x-th type of firm, and then obtain the type dissimilarity D between the new firm and the x-th type of firm. x D x The average dissimilarity of N features between the new enterprise and enterprises of type x is used. Substituting each value into x=1,2,…,X, we obtain the type dissimilarity between the new enterprise and enterprises of type X. The enterprise type with the lowest type dissimilarity is selected as the type of the new enterprise, thus classifying it as the type with the lowest dissimilarity. For the consulting scenario of the new enterprise, its enterprise data is processed using a standardized process consistent with that of historical enterprises to ensure the comparability and effectiveness of the extracted enterprise features. By comprehensively comparing and analyzing the features of the new enterprise with the standard features of various types of enterprises in the database, the optimal matching type is selected based on the dissimilarity, achieving accurate classification of the new enterprise. This avoids the mismatch problem caused by the single dimension and inconsistent standards in traditional classification, allowing the new enterprise to accurately connect with the experience system of similar enterprises, laying a solid foundation for subsequent targeted information recommendations, further improving the accuracy of consulting matching, reducing interference from irrelevant information, and helping enterprises quickly obtain consulting directions that meet their needs.

[0008] Furthermore, in step S3, a keyword database is preset. When a new enterprise seeks information, keywords are retrieved from the information inquired about by the new enterprise, and historical search records of similar enterprises are retrieved to obtain Y search results for similar enterprises. For the y-th search result, y=1,2,…,Y, the number of times the similar enterprise retrieves the search result is P. After retrieving the y-th search result P times, the search satisfaction of the similar enterprise after the p-th retrieval of the y-th search result is E. p A rating system, such as a 1-5 rating, can be set after the search results are obtained. This rating is used as the search satisfaction level, thus obtaining the search satisfaction level for obtaining the y-th search result P times. The search satisfaction level for obtaining the y-th search result P times is {E1, E2, ..., E...} p ,…,E P}, where search satisfaction E p Corresponding company F p Company F pThe firm characteristics are {G1, G2, ..., G...} n ,…,G N}, where G n Company F p The nth firm characteristic is used to calculate the new firm and firm F. p Enterprise similarity H p : ; Substitute p = 1, 2, ..., P one by one to obtain the enterprise similarity between the new enterprise and the P existing enterprises, and then obtain the search satisfaction ratio J of the p-th search result obtained for the y-th search result. p J p For H p The ratio of the sum of the similarities between the new enterprise and the P existing enterprises is used as the input for each of the P similarities to obtain the y-th search result. This yields the percentage coefficient of search satisfaction for the y-th search result obtained in P instances, and thus the recommendation coefficient K for the y-th search result. y K y Let E be the sum of the products of the search satisfaction rate and the corresponding percentage coefficient for each of the P times the y-th search result is obtained, where the search satisfaction rate E is... p With proportion coefficient J p Correspondingly, this approach precisely extracts the core consulting needs of new businesses through a pre-set keyword database. Simultaneously, it deeply analyzes historical search records and satisfaction feedback from similar businesses, incorporating the similarity between the new business and its peers into the recommendation coefficient calculation. This breaks through the traditional single-recommendation logic that relies solely on search volume. By analyzing the correlation between satisfaction and business similarity, the recommendation coefficient better reflects the suitability of search results for the new business. It builds upon the foundation of precise business classification mentioned earlier and further refines the recommendation criteria, effectively avoiding information redundancy caused by generalized recommendations. This makes search result recommendations more targeted and practical, helping businesses quickly identify high-quality information that meets their needs and continuously improving the accuracy and efficiency of information matching.

[0009] Furthermore, in step S4, by substituting y=1,2,…,Y one by one, the recommendation coefficients of the Y search results are obtained, and the recommendation coefficients of the Y search results are {K1,K2,…,K...}. y ,…,K Y} Select the Q search results with the highest recommendation coefficients as candidate search results, where Q is the preset number of candidate search results. Number the Q candidate search results as {L1, L2, ..., L}. y ,…,L Y The algorithm retrieves the top I search keywords for a new company, with the current time as the endpoint. It then analyzes the relevance of these top I keywords to the y-th search result. For the ith search keyword, it retrieves the search records of similar companies, assuming that similar companies search for the ith keyword an O time interval. iWhen searching for the i-th keyword among similar companies, the number of times the α-th search result contains the y-th search result is R. i The nearest α search results include: the search result obtained from searching for the i-th search keyword, the search results obtained from (α-1) / 2 searches before searching for the i-th search keyword, and the search results obtained from (α-1) / 2 searches after searching for the i-th search keyword, thereby obtaining the relevance S of the y-th search result for the i-th search keyword. i_y S i_y =R i / O i Substitute each i = 1, 2, ..., I to obtain the relevance of the y-th search result for the first I search keywords. The relevance of the first I search keywords to the y-th search result is denoted as {S}. 1_y ,S 2_y ,…,S i_y ,…,S I_y}, and then obtain the comprehensive recommendation coefficient T of the y-th search result. y T y The average relevance of the first I search keywords to the y-th search result is multiplied by U1 and K. y The search results are multiplied by U1, where U1 is the weight of the relevance of the first I search keywords to the y-th search result, and U2 is the weight of the recommendation coefficient. The relevance weight and recommendation weight are initially set to 1. Search results are then recommended to new businesses from highest to lowest based on the overall recommendation coefficient. High-recommendation-coefficient candidate search results are filtered, focusing on high-quality candidates. This is further combined with the new business's historical search keywords, and by analyzing the search behavior of similar businesses, the potential relevance between keywords and candidate results is analyzed, making the relevance assessment more aligned with actual search scenarios. Simultaneously, a weighting mechanism is introduced to comprehensively consider the recommendation coefficient and keyword relevance, constructing a scientific comprehensive recommendation system. This builds upon the previous precise classification and recommendation coefficient calculation, further refining the recommendation logic. This effectively avoids the one-sidedness of single-dimensional recommendations, making the matching of search results with the new business's consultation needs more accurate, helping businesses quickly filter high-quality information, and significantly improving the efficiency and suitability of information consultation.

[0010] Furthermore, in step S5, before recommending search results, the search results with the highest recommendation coefficient and the search results with the highest relevance are marked. The average search satisfaction rate of the search result with the highest recommendation coefficient in historical data is V1, and the average search satisfaction rate of the search result with the highest relevance in historical data is V2. The relevance weight is then adjusted to U1*V. 1 / (V 1+ V2), adjust the recommendation weight to U1*V 2 / (V 1+V2); Core search results are marked before recommendations, and weight configurations are dynamically adjusted based on historical satisfaction data for two key results, breaking the rigid limitations of traditional fixed weights. By using historical satisfaction as the core basis for weight adjustment, the configuration of association and recommendation proportions is made more aligned with actual usage effects, making the calculation of the comprehensive recommendation coefficient more scientific and adaptable. This approach builds upon the previous sections on precise classification, association analysis, and recommendation system construction, further optimizing the recommendation logic and ensuring that the comprehensive recommendation results both consider the high-quality foundation of the recommendation coefficient and highlight the demand relevance of keyword associations, effectively improving the accuracy of consultation matching. Simultaneously, the dynamic weight mechanism can respond to historical usage feedback in real time, continuously optimizing the recommendation direction, reducing interference from irrelevant information, and helping enterprises more efficiently obtain high-quality information that matches their core needs, steadily improving the practical value of information consultation and user experience.

[0011] Furthermore, in step S6, after accumulating β keyword searches, where β is a preset keyword search threshold for database updates, the database is updated in real time.

[0012] An AI-based intelligent optimization system for information consulting strategies includes: a historical enterprise classification database module, a new enterprise feature comparison and segmentation module, a preset thesaurus initial selection and recommendation module, a historical keyword comprehensive recommendation module, a result labeling weight configuration module, and a cumulative frequency real-time update module. The historical enterprise classification and database building module is used to classify enterprises based on historical data and establish an enterprise characteristic database; The new enterprise feature comparison and analysis module is used to compare the enterprise features of the new enterprise with the enterprise features in the enterprise feature database when receiving information inquiries from a new enterprise, in order to obtain the enterprise type of the new enterprise. The preset keyword database initial recommendation module is used to preset a keyword database. When a new enterprise seeks information, it considers the similarity between the new enterprise and the enterprises in the database to obtain the initial recommendation level of the search results obtained by the enterprises in the database. The historical keyword comprehensive recommendation module is used to retrieve keywords from the new enterprise's historical searches, consider the relevance between the historical search keywords and the search results, and comprehensively consider the relevance between the keywords and the search results and the initial recommendation of the search results to provide search results for the new enterprise. The result labeling weight configuration module is used to label the search results before recommending them, taking into account the historical satisfaction of the search results, and setting weights for relevance and initial recommendation. The cumulative search count update module is used to accumulate keyword search counts and update the database in real time.

[0013] Compared with existing technologies, the beneficial effects achieved by this invention are as follows: On the one hand, through systematic processing and feature extraction of enterprise data, combined with feature interval division, accurate classification of enterprises is achieved, constructing a comprehensive enterprise feature database. When consulting with new enterprises, the type is determined through detailed comparison of enterprise features, ensuring that subsequent recommendations are based on the effective experience and common needs of similar enterprises. Simultaneously, by combining satisfaction feedback from similar enterprises regarding search results, and similarity analysis between new enterprises and similar enterprises, search result recommendations are made more aligned with the actual business scenarios and consulting needs of new enterprises, avoiding interference from irrelevant information. This targeted recommendation mechanism significantly improves the matching efficiency of information consulting, helps enterprises quickly obtain valuable consulting results, reduces information filtering time, and significantly optimizes the enterprise's information consulting experience.

[0014] On the one hand, by introducing relevance weighting and recommendation weighting, and analyzing the relevance between historical search keywords and search results for new enterprises, the weighting configuration is dynamically adjusted based on historical satisfaction data, breaking the rigid limitations of traditional fixed recommendation models. By comprehensively considering the recommendation coefficient and keyword relevance, a scientific comprehensive recommendation coefficient calculation system is constructed, enabling the recommendation strategy to accurately adapt to the dynamic changes in enterprise consulting needs. Furthermore, by analyzing the relevance of search results in adjacent search behaviors of similar enterprises, the accuracy of keyword and search result correlation analysis is further improved, making the recommendation strategy more intelligent and flexible, effectively solving the problem that traditional consulting strategies are unable to adapt to the personalized and dynamic needs of enterprises.

[0015] On the other hand, a real-time database update mechanism is implemented to continuously optimize the enterprise characteristic database and search result data as keyword search volume accumulates, ensuring the timeliness and comprehensiveness of the data. By continuously incorporating new enterprise data, search records, and satisfaction feedback, core aspects such as enterprise classification standards, recommendation coefficient calculation, and correlation analysis are continuously optimized, ensuring the system operates efficiently in the long term. Simultaneously, based on massive historical data and real-time data updates, the system can continuously improve the accuracy of enterprise classification, the adaptability of recommendation strategies, and enhance the stability and reliability of consulting services. This provides enterprises with long-term, continuous, high-quality information consulting support, helping them efficiently resolve various consulting needs during business development. Attached Figure Description

[0016] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings: Figure 1 This is a structural diagram of an intelligent optimization system for information consulting strategies based on artificial intelligence, according to the present invention. Figure 2This is a flowchart of an intelligent optimization method for information consulting strategies based on artificial intelligence, according to the present invention. Detailed Implementation

[0017] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. 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 are within the scope of protection of the present invention.

[0018] Please see Figure 1 and Figure 2 This invention provides a technical solution: an intelligent optimization method for information consulting strategies based on artificial intelligence, comprising the following steps: S1. Classify enterprises based on historical data and establish an enterprise characteristic database; S2. When receiving information inquiries from a new enterprise, compare the new enterprise's characteristics with those in the enterprise characteristic database to obtain the new enterprise's enterprise type; S3. Pre-set keyword database: When a new company seeks information, the similarity between the new company and companies in the database is considered to obtain the preliminary recommendation level of the search results obtained by the companies in the database. S4. Retrieve keywords from the new company's historical searches, consider the relevance between the keywords and search results, and comprehensively consider the relevance between the keywords and search results and the initial recommendation of the search results to provide search results for the new company. S5. Before recommending search results, mark the search results, consider the historical satisfaction of the search results, and set weights for relevance and initial recommendation. S6. Cumulative keyword search counts, with real-time database updates.

[0019] After filtering the search results, the historical search keywords of the new company are retrieved, and the relevance between the keywords and the search results is analyzed. During the analysis, the search records of similar companies are called up, and the number of times the target result is included in the search results when similar companies use the corresponding keywords is statistically analyzed. The relevance between a single keyword and its corresponding result is determined, and then the overall relevance level is obtained by combining the relevance of all relevant keywords.

[0020] In step S1, after authorization, enterprise data is acquired based on historical data. This data is preprocessed and normalized to obtain enterprise characteristics. These characteristics are then divided into characteristic intervals, and enterprises are categorized into X types, resulting in a database of enterprise characteristics for each category. Through in-depth mining of historical enterprise data after authorization, invalid information is removed through standardization, core characteristics are extracted, and a database of enterprise characteristics is constructed by scientifically dividing characteristic intervals, achieving systematic and precise enterprise classification. When consulting with new enterprises, classification is completed by accurately comparing their own characteristics with the classification standards in the database, ensuring that subsequent information recommendations strictly anchor to the effective experience and common needs of similar enterprises. This significantly reduces the interference of irrelevant information on the consultation results, making the information consultation direction more aligned with the actual business scenarios and core needs of enterprises. It helps enterprises quickly filter and obtain suitable key information, effectively improving the efficiency and accuracy of information consultation and laying a solid foundation for subsequent intelligent recommendations throughout the entire process.

[0021] In step S2, upon receiving an information inquiry from a new enterprise, the enterprise data of the new enterprise is obtained. After preprocessing and normalizing the enterprise data, the enterprise characteristics of the new enterprise are obtained. The enterprise characteristics of the new enterprise are {A1, A2, ..., A...} n ,…,A N}, where N represents the number of firm features, A n This represents the nth enterprise characteristic of a new enterprise. The enterprise characteristic database is accessed, and the standard enterprise characteristic for the xth type of enterprise is {B1_ x B2_ x ,…,B n _ x ,…,B N _ x}, where B n _ x B represents the nth standard firm characteristic of firm x. n Let C be the average of the nth firm characteristic of all firms in the x-th firm category, and then we can obtain the dissimilarity C between the nth characteristic of the new firm and the firms in the x-th firm category. n _ x C n_x =|(A n -B n _ x ) / B n _ x Substitute each of the following into n=1,2,…,N to obtain the dissimilarity of N features between the new firm and the x-th type of firm, and then obtain the type dissimilarity D between the new firm and the x-th type of firm. x D xThe average dissimilarity of N features between the new enterprise and enterprises of type x is used. Substituting each value into x=1,2,…,X, we obtain the type dissimilarity between the new enterprise and enterprises of type X. The enterprise type with the lowest type dissimilarity is selected as the type of the new enterprise, thus classifying it as the type with the lowest dissimilarity. For the consulting scenario of the new enterprise, its enterprise data is processed using a standardized process consistent with that of historical enterprises to ensure the comparability and effectiveness of the extracted enterprise features. By comprehensively comparing and analyzing the features of the new enterprise with the standard features of various types of enterprises in the database, the optimal matching type is selected based on the dissimilarity, achieving accurate classification of the new enterprise. This avoids the mismatch problem caused by the single dimension and inconsistent standards in traditional classification, allowing the new enterprise to accurately connect with the experience system of similar enterprises, laying a solid foundation for subsequent targeted information recommendations, further improving the accuracy of consulting matching, reducing interference from irrelevant information, and helping enterprises quickly obtain consulting directions that meet their needs.

[0022] In step S3, a keyword database is preset. When a new company inquires about information, keywords are retrieved from the information inquired about by the new company. Historical search records of similar companies are retrieved to obtain Y search results for similar companies. For the y-th search result, y=1,2,…,Y, the number of times the similar company retrieves the search result is P. After retrieving the y-th search result P times, the search satisfaction of the similar company after the p-th retrieval of the y-th search result is E. p A rating system, such as a 1-5 rating, can be set after the search results are obtained. This rating is used as the search satisfaction level, thus obtaining the search satisfaction level for obtaining the y-th search result P times. The search satisfaction level for obtaining the y-th search result P times is {E1, E2, ..., E...} p ,…,E P}, where search satisfaction E p Corresponding company F p Company F p The firm characteristics are {G1, G2, ..., G...} n ,…,G N}, where G n Company F p The nth firm characteristic is used to calculate the new firm and firm F. p Enterprise similarity H p : ; Substitute p = 1, 2, ..., P one by one to obtain the enterprise similarity between the new enterprise and the P existing enterprises, and then obtain the search satisfaction ratio J of the p-th search result obtained for the y-th search result. p J p For H pThe ratio of the sum of the similarities between the new enterprise and the P existing enterprises is used as the input for each of the P similarities to obtain the y-th search result. This yields the percentage coefficient of search satisfaction for the y-th search result obtained in P instances, and thus the recommendation coefficient K for the y-th search result. y K y Let E be the sum of the products of the search satisfaction rate and the corresponding percentage coefficient for each of the P times the y-th search result is obtained, where the search satisfaction rate E is... p With proportion coefficient J p Correspondingly, this approach precisely extracts the core consulting needs of new businesses through a pre-set keyword database. Simultaneously, it deeply analyzes historical search records and satisfaction feedback from similar businesses, incorporating the similarity between the new business and its peers into the recommendation coefficient calculation. This breaks through the traditional single-recommendation logic that relies solely on search volume. By analyzing the correlation between satisfaction and business similarity, the recommendation coefficient better reflects the suitability of search results for the new business. It builds upon the foundation of precise business classification mentioned earlier and further refines the recommendation criteria, effectively avoiding information redundancy caused by generalized recommendations. This makes search result recommendations more targeted and practical, helping businesses quickly identify high-quality information that meets their needs and continuously improving the accuracy and efficiency of information matching.

[0023] In step S4, substitute y=1,2,…,Y one by one to obtain the recommendation coefficients of the Y search results. The recommendation coefficients of the Y search results are {K1,K2,…,K...} y ,…,K Y} Select the Q search results with the highest recommendation coefficients as candidate search results, where Q is the preset number of candidate search results. Number the Q candidate search results as {L1, L2, ..., L}. y ,…,L Y The algorithm retrieves the top I search keywords for a new company, with the current time as the endpoint. It then analyzes the relevance of these top I keywords to the y-th search result. For the ith search keyword, it retrieves the search records of similar companies, assuming that similar companies search for the ith keyword an O time interval. i When searching for the i-th keyword among similar companies, the number of times the α-th search result contains the y-th search result is R. i The nearest α search results include: the search result obtained from searching for the i-th search keyword, the search results obtained from (α-1) / 2 searches before searching for the i-th search keyword, and the search results obtained from (α-1) / 2 searches after searching for the i-th search keyword, thereby obtaining the relevance S of the y-th search result for the i-th search keyword. i_y S i_y =R i / O iSubstitute each i = 1, 2, ..., I to obtain the relevance of the y-th search result for the first I search keywords. The relevance of the first I search keywords to the y-th search result is denoted as {S}. 1_y ,S 2_y ,…,S i_y ,…,S I_y}, and then obtain the comprehensive recommendation coefficient T of the y-th search result. y T y The average relevance of the first I search keywords to the y-th search result is multiplied by U1 and K. y The search results are multiplied by U1, where U1 is the weight of the relevance of the first I search keywords to the y-th search result, and U2 is the weight of the recommendation coefficient. The relevance weight and recommendation weight are initially set to 1. Search results are then recommended to new businesses from highest to lowest based on the overall recommendation coefficient. High-recommendation-coefficient candidate search results are filtered, focusing on high-quality candidates. This is further combined with the new business's historical search keywords, and by analyzing the search behavior of similar businesses, the potential relevance between keywords and candidate results is analyzed, making the relevance assessment more aligned with actual search scenarios. Simultaneously, a weighting mechanism is introduced to comprehensively consider the recommendation coefficient and keyword relevance, constructing a scientific comprehensive recommendation system. This builds upon the previous precise classification and recommendation coefficient calculation, further refining the recommendation logic. This effectively avoids the one-sidedness of single-dimensional recommendations, making the matching of search results with the new business's consultation needs more accurate, helping businesses quickly filter high-quality information, and significantly improving the efficiency and suitability of information consultation.

[0024] In step S5, before recommending search results, the search results with the highest recommendation coefficient and the search results with the highest relevance are marked. The average search satisfaction of the search result with the highest recommendation coefficient in historical data is V1, and the average search satisfaction of the search result with the highest relevance in historical data is V2. The relevance weight is then adjusted to U1*V. 1 / (V 1+ V2), adjust the recommendation weight to U1*V 2 / (V 1+V2); Core search results are marked before recommendations, and weight configurations are dynamically adjusted based on historical satisfaction data for two key results, breaking the rigid limitations of traditional fixed weights. By using historical satisfaction as the core basis for weight adjustment, the configuration of association and recommendation proportions is made more aligned with actual usage effects, making the calculation of the comprehensive recommendation coefficient more scientific and adaptable. This approach builds upon the previous sections on precise classification, association analysis, and recommendation system construction, further optimizing the recommendation logic and ensuring that the comprehensive recommendation results both consider the high-quality foundation of the recommendation coefficient and highlight the demand relevance of keyword associations, effectively improving the accuracy of consultation matching. Simultaneously, the dynamic weight mechanism can respond to historical usage feedback in real time, continuously optimizing the recommendation direction, reducing interference from irrelevant information, and helping enterprises more efficiently obtain high-quality information that matches their core needs, steadily improving the practical value of information consultation and user experience.

[0025] In step S6, after accumulating β keyword searches, where β is a preset keyword search threshold for database updates, the database is updated in real time.

[0026] An AI-based intelligent optimization system for information consulting strategies includes: a historical enterprise classification database module, a new enterprise feature comparison and analysis module, a preset thesaurus initial selection and recommendation module, a historical keyword comprehensive recommendation module, a result labeling weight configuration module, and a cumulative frequency real-time update module. The historical enterprise classification and database building module is used to classify enterprises based on historical data and establish an enterprise characteristic database; The new enterprise feature comparison and analysis module is used to compare the new enterprise's enterprise features with the enterprise features in the enterprise feature database when receiving information inquiries about a new enterprise, in order to obtain the enterprise type of the new enterprise; The preset keyword database and initial recommendation module is used to preset the keyword database. When a new company seeks information, it considers the similarity between the new company and companies in the database to obtain the initial recommendation score of the search results obtained by the companies in the database. The historical keyword comprehensive recommendation module is used to retrieve keywords from the new enterprise's historical searches. It considers the relevance between the historical search keywords and the search results, and comprehensively considers the relevance between the keywords and the search results and the initial recommendation of the search results to provide search results for the new enterprise. The result labeling weight configuration module is used to label search results before recommending them, taking into account the historical satisfaction of the search results and setting weights for relevance and initial recommendation. The cumulative search count update module is used to accumulate keyword search counts and update the database in real time.

[0027] Example 1: First, with authorization, historical enterprise data is deeply mined. The data undergoes preprocessing and normalization to systematically remove invalid and redundant information, accurately extracting the core characteristics of each enterprise. Based on these extracted characteristics, a scientific feature interval system is constructed, and enterprises are systematically categorized accordingly, forming a clearly classified and well-defined enterprise feature database. This step, through standardization and refined classification, lays the foundation for accurate enterprise matching, ensuring the comprehensiveness and scientific rigor of the enterprise classification system upon which subsequent recommendations are based.

[0028] When receiving inquiries about new businesses, the data is processed using the same standardized procedures as for historical businesses, ensuring the comparability and effectiveness of extracted business characteristics. Subsequently, the business characteristic database is accessed, and the core characteristics of the new business are comprehensively compared and analyzed against the standard characteristics of various businesses in the database. Based on the degree of characteristic fit, the optimal matching type is selected, achieving accurate categorization of the new business. This allows new businesses to accurately connect with the experience systems of similar companies, ensuring the suitability of the consultation direction from the outset and avoiding inaccurate recommendations caused by the biases of traditional categorization.

[0029] A pre-set keyword database is used to accurately capture the core needs of new business inquiries, while also deeply mining historical search records and satisfaction feedback from similar companies. Combining the similarity of features between new and similar businesses, a multi-dimensional recommendation coefficient calculation system is constructed. This breaks through the traditional single recommendation logic that relies solely on search frequency, enabling the recommendation coefficient to more accurately reflect the actual suitability of search results for new businesses and reducing information redundancy caused by generalized recommendations.

[0030] High-quality search results with high recommendation coefficients are selected as the candidate range. Then, recent historical search keywords for new companies are retrieved. By analyzing the search behavior of similar companies in the vicinity, the potential correlation between keywords and candidate results is analyzed. A dual-weighting mechanism of correlation ratio and recommendation ratio is introduced to construct a comprehensive recommendation coefficient system. Preliminary recommendation results are generated by sorting the results according to their coefficients, avoiding the bias of single-dimensional recommendations and improving matching accuracy.

[0031] Prior to recommendations, core search results with high recommendation coefficients and high relevance are marked, and their weights are dynamically adjusted based on historical satisfaction data. This ensures that the weight settings better reflect actual usage, guaranteeing that the overall recommendation results balance high-quality fundamentals with a strong fit for the needs of new businesses, and continuously optimizing the recommendation direction.

[0032] Finally, a real-time database update mechanism is established. As keyword search behavior continues to accumulate, key data such as the enterprise characteristic database, search records, and satisfaction feedback are dynamically updated. This ensures that enterprise classification standards and recommendation logic always rely on the latest data, guaranteeing the long-term effectiveness of the system and providing enterprises with continuous and stable high-quality information consulting support.

[0033] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary sensing device embodiments described above, and that the invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the invention. Therefore, the embodiments should be considered in all respects as exemplary and non-limiting, and the scope of the invention is defined by the appended claims rather than the foregoing description. Thus, all variations falling within the meaning and scope of equivalents of the claims are intended to be included within the present invention. No reference numerals in the claims should be construed as limiting the scope of the claims.

Claims

1. An intelligent optimization method for information consulting strategies based on artificial intelligence, characterized in that: The method includes the following steps: S1. Classify enterprises based on historical data and establish an enterprise characteristic database; S2. When receiving information inquiries from a new enterprise, compare the new enterprise's characteristics with those in the enterprise characteristic database to obtain the new enterprise's enterprise type; S3. Pre-set keyword database: When a new company seeks information, the similarity between the new company and companies in the database is considered to obtain the preliminary recommendation level of the search results obtained by the companies in the database. S4. Retrieve keywords from the new company's historical searches, consider the relevance between the keywords and search results, and comprehensively consider the relevance between the keywords and search results and the initial recommendation of the search results to provide search results for the new company. S5. Before recommending search results, mark the search results, consider the historical satisfaction of the search results, and set weights for relevance and initial recommendation. S6. Cumulative keyword search counts, with real-time database updates.

2. The intelligent optimization method for information consulting strategies based on artificial intelligence according to claim 1, characterized in that: After filtering the search results, the historical search keywords of the new company are retrieved, and the relevance between the keywords and the search results is analyzed. During the analysis, the search records of similar companies are called up, and the number of times the target result is included in the search results when similar companies use the corresponding keywords is statistically analyzed. The relevance between a single keyword and its corresponding result is determined, and then the overall relevance level is obtained by combining the relevance of all relevant keywords.

3. The intelligent optimization method for information consulting strategies based on artificial intelligence according to claim 2, characterized in that: In step S1, after authorization, enterprise data is obtained based on historical data. After preprocessing and normalizing the enterprise data, enterprise characteristics are obtained, and feature intervals are divided for the enterprise characteristics. Enterprises are divided into X types based on their characteristic intervals, thus obtaining a database of enterprise characteristics for each X type of enterprise.

4. The intelligent optimization method for information consulting strategies based on artificial intelligence according to claim 3, characterized in that: In step S2, upon receiving an information inquiry from a new enterprise, the enterprise data of the new enterprise is obtained. After preprocessing and normalizing the enterprise data, the enterprise characteristics of the new enterprise are obtained. The enterprise characteristics of the new enterprise are {A1, A2, ..., A...} n ,…,A N }, where N represents the number of firm features, A n This represents the nth enterprise characteristic of a new enterprise. The enterprise characteristic database is accessed, and the standard enterprise characteristic for the xth type of enterprise is {B1_ x B2_ x ,…,B n _ x ,…,B N _ x }, where B n _ x B represents the nth standard firm characteristic of firm x. n Let C be the average of the nth firm characteristic of all firms in the x-th firm category, and then we can obtain the dissimilarity C between the nth characteristic of the new firm and the firms in the x-th firm category. n _ x C n_x =|(A n -B n _ x ) / B n _ x Substitute each of the following into n=1,2,…,N to obtain the dissimilarity of N features between the new firm and the x-th type of firm, and then obtain the type dissimilarity D between the new firm and the x-th type of firm. x D x Let x be the average of the N characteristics of the new enterprise and the x-th type of enterprise. Substitute each of x=1,2,…,X to obtain the type dissimilarity between the new enterprise and the x-th type of enterprise. Select the enterprise type with the lowest type dissimilarity as the type of the new enterprise and classify the new enterprise into the enterprise type with the lowest enterprise dissimilarity.

5. The intelligent optimization method for information consulting strategies based on artificial intelligence according to claim 4, characterized in that: In step S3, a keyword database is preset. When a new company inquires about information, keywords are retrieved from the information inquired about by the new company. Historical search records of similar companies are retrieved to obtain Y search results for similar companies. For the y-th search result, y=1,2,…,Y, the number of times the similar company retrieves the search result is P. After retrieving the y-th search result P times, the search satisfaction of the similar company after the p-th retrieval of the y-th search result is E. p This leads to the search satisfaction rate for obtaining the y-th search result P times, where the search satisfaction rate for obtaining the y-th search result P times is {E1, E2, ..., E...}. p ,…,E P }, where search satisfaction E p Corresponding company F p Company F p The firm characteristics are {G1, G2, ..., G...} n ,…,G N }, where G n Company F p The nth firm characteristic is used to calculate the new firm and firm F. p Enterprise similarity H p : ； Substitute p = 1, 2, ..., P one by one to obtain the enterprise similarity between the new enterprise and the P existing enterprises, and then obtain the search satisfaction ratio J of the p-th search result obtained for the y-th search result. p J p For H p The ratio of the sum of the similarities between the new enterprise and the P existing enterprises is used as the input for each of the P similarities to obtain the y-th search result. This yields the percentage coefficient of search satisfaction for the y-th search result obtained in P instances, and thus the recommendation coefficient K for the y-th search result. y K y Let E be the sum of the products of the search satisfaction rate and the corresponding percentage coefficient for each of the P times the y-th search result is obtained, where the search satisfaction rate E is... p With proportion coefficient J p correspond.

6. The intelligent optimization method for information consulting strategies based on artificial intelligence according to claim 5, characterized in that: In step S4, substitute y=1,2,…,Y one by one to obtain the recommendation coefficients of the Y search results. The recommendation coefficients of the Y search results are {K1,K2,…,K...} y ,…,K Y } Select the Q search results with the highest recommendation coefficients as candidate search results, where Q is the preset number of candidate search results. Number the Q candidate search results as {L1, L2, ..., L}. y ,…,L Y The algorithm retrieves the top I search keywords for a new company, with the current time as the endpoint. It then analyzes the relevance of these top I keywords to the y-th search result. For the ith search keyword, it retrieves the search records of similar companies, assuming that similar companies search for the ith keyword an O time interval. i When searching for the i-th keyword among similar companies, the number of times the α-th search result contains the y-th search result is R. i The nearest α search results include: the search result obtained from searching for the i-th search keyword, the search results obtained from (α-1) / 2 searches before searching for the i-th search keyword, and the search results obtained from (α-1) / 2 searches after searching for the i-th search keyword, thereby obtaining the relevance S of the y-th search result for the i-th search keyword. i_y S i_y =R i / O i Substitute each i = 1, 2, ..., I to obtain the relevance of the y-th search result for the first I search keywords. The relevance of the first I search keywords to the y-th search result is denoted as {S}. 1_y ,S 2_y ,…,S i_y ,…,S I_y }, and then obtain the comprehensive recommendation coefficient T of the y-th search result. y T y The average relevance of the first I search keywords to the y-th search result is multiplied by U1 and K. y Multiply by U1 and U2, where U1 is the relevance weight of the first I search keywords to the y-th search result, and U2 is the recommendation weight of the recommendation coefficient. The relevance weight and recommendation weight are initially set to 1. Then, the search results are recommended to the new enterprise from high to low according to the comprehensive recommendation coefficient.

7. The intelligent optimization method for information consulting strategies based on artificial intelligence according to claim 6, characterized in that: In step S5, before recommending search results, the search results with the highest recommendation coefficient and the search results with the highest relevance are marked. The average search satisfaction of the search result with the highest recommendation coefficient in historical data is V1, and the average search satisfaction of the search result with the highest relevance in historical data is V2. The relevance weight is then adjusted to U1*V. 1 / (V 1+ V2), adjust the recommendation weight to U1*V 2 / (V 1+ V2).

8. The intelligent optimization method for information consulting strategies based on artificial intelligence according to claim 6, characterized in that: In step S6, after accumulating β keyword searches, where β is a preset keyword search threshold for database updates, the database is updated in real time.

9. An intelligent optimization system for information consulting strategies based on artificial intelligence, wherein the system is applied to the intelligent optimization method for information consulting strategies based on artificial intelligence as described in any one of claims 1-8, characterized in that: The system includes: a historical enterprise classification and database building module, a new enterprise feature comparison and typing module, a preset thesaurus initial selection and recommendation module, a historical keyword comprehensive recommendation module, a result labeling weight configuration module, and a cumulative number of real-time updates module. The historical enterprise classification and database building module is used to classify enterprises based on historical data and establish an enterprise characteristic database; The new enterprise feature comparison and analysis module is used to compare the enterprise features of the new enterprise with the enterprise features in the enterprise feature database when receiving information inquiries from a new enterprise, in order to obtain the enterprise type of the new enterprise. The preset keyword database initial recommendation module is used to preset a keyword database. When a new enterprise seeks information, it considers the similarity between the new enterprise and the enterprises in the database to obtain the initial recommendation level of the search results obtained by the enterprises in the database. The historical keyword comprehensive recommendation module is used to retrieve keywords from the new enterprise's historical searches, consider the relevance between the historical search keywords and the search results, and comprehensively consider the relevance between the keywords and the search results and the initial recommendation of the search results to provide search results for the new enterprise. The result labeling weight configuration module is used to label the search results before recommending them, taking into account the historical satisfaction of the search results, and setting weights for relevance and initial recommendation. The cumulative search count update module is used to accumulate keyword search counts and update the database in real time.

Images