Business opportunity recommendation system and method based on collaborative filtering technology

Abstract

The invention discloses a business opportunity recommendation system and method based on a collaborative filtering technology. The system comprises a business opportunity data collection and processing module, a user behavior data collection and processing module, a collaborative filtering recommendation engine, a recommendation result generation and display module, a feedback optimization module and a data storage and management module. By using a collaborative filtering technology, the system can carry out personalized recommendation according to user historical behaviors and business opportunity data, continuously optimize recommendation results according to user feedback, and is suitable for various business environments and application scenes; the defect that a traditional rule-based or content-based recommendation system is difficult to fully consider dynamic changes of user preferences and behaviors is effectively overcome.

Description

Technical Field

[0001] The present invention belongs to the technical field of big data analysis and processing, and specifically relates to a business opportunity capability recommendation system and method based on collaborative filtering technology. Background Art

[0002] In the business environment, effective business opportunity management and accurate capability recommendation are crucial to the development of enterprises. Traditional business opportunity capability recommendation methods mainly rely on rule-based or content-based recommendation systems, but these methods often fail to fully consider the dynamic changes of user preferences and historical behaviors, and the recommendation results lack personalization and accuracy. Rule-based recommendation systems rely on pre-set rules, which are often static and cannot adapt to changes in user needs; while content-based recommendation systems can make recommendations based on users' past behaviors, but their performance is often unsatisfactory when faced with sparse data or diversified user preferences. With the development of the Internet and big data technology, recommendation systems based on collaborative filtering have gradually attracted attention because they can use user historical behavior data for personalized recommendations. Collaborative filtering technology can provide more personalized and accurate recommendations by analyzing the similarities between users and user historical behavior data. Therefore, it is necessary to design a business opportunity capability recommendation system and method based on collaborative filtering technology to solve the problems of lack of personalization and insufficient accuracy in traditional business opportunity capability recommendation methods.

[0003] The following are the public Chinese patent documents related to collaborative filtering, business opportunity capability recommendation system and method:

[0004] Patent application number CN202310363472.2 discloses a business opportunity recommendation method, device and electronic device, the method comprising: obtaining a business opportunity customer list, calculating the business opportunity value rate of each business opportunity customer in the business opportunity customer list, and generating an initial business opportunity value rate for each business opportunity customer; screening each business opportunity customer based on the initial business opportunity value rate to obtain target customers whose initial business opportunity value rate meets predetermined conditions; obtaining real-time customer data of the target customers, inputting the real-time customer data into a business opportunity value rate identification model to obtain a target business opportunity value rate for the target customers; and generating product recommendation data corresponding to the target customers based on the target business opportunity value rate.

[0005] Patent application number CN202410390773.9 discloses a diversified recommendation method based on improved collaborative filtering and penalty label scoring, including: obtaining relevant data of the recommendation object and constructing a user-item interaction matrix; calculating user similarity and item similarity according to the user-item interaction matrix, and generating a user-based recommendation list and an item-based recommendation list; fusing the user-based recommendation list and the item-based recommendation list based on preset weights to obtain a fused recommendation list; collecting item labels, using a label scoring algorithm to calculate the label popularity and reducing the scores of items in the fused recommendation list according to the label popularity to generate a final Top-K recommendation list.

[0006] Patent publication number CN111460271A discloses an intelligent personalized recommendation algorithm for network information. The implementation of the recommendation algorithm includes the following steps: Step 1: Collect user preference information: The system autonomously provides users with multiple ways to display their preference information, and obtains preference data based on specific user behavior; Step 2: Preference data preprocessing: In-depth analysis and processing of the obtained preference data where accurate information cannot be directly obtained; Step 3: Find similar users or items: After analyzing the user's preferences, similarity calculation is performed according to the corresponding preferences, including user-based collaborative filtering, project-based collaborative filtering, and content-based filtering algorithms; Step 4: A hybrid recommendation algorithm system is formed by mixing multiple recommendation algorithms. This patent mines and generates user portraits based on the preferences of different users, provides each user with personalized recommendation content with "thousands of faces", helps media, e-commerce and other industries effectively improve click-through rate, conversion rate and user stickiness, greatly increases customer operating benefits, and can provide personalized recommendation results for each user on the homepage, channel page, guess you like page, discovery page, etc. of the APP. However, this patent only analyzes historical data, but does not collect real-time feedback from users to optimize the model and algorithm.

[0007] Patent publication number CN115705340A discloses a collaborative filtering recommendation method, which includes: obtaining user historical behavior data from a big data platform and storing it in a distributed file system; performing similarity measurement and multi-stage parallel processing on the user historical behavior data stored in the distributed file system to generate a set of items to be recommended; and recommending items in the set of items to be recommended to target users. This patent also does not take into account user feedback information, and the accuracy and effectiveness of the recommended content needs to be further improved. Summary of the invention

[0008] The purpose of the present invention is to address the deficiencies of the prior art and provide a business opportunity capability recommendation system and method based on collaborative filtering technology. By utilizing collaborative filtering technology, the present invention enables the system to make personalized recommendations based on user historical behavior and business opportunity data, and continuously optimizes the recommendation results based on user feedback, which is applicable to a variety of business environments and application scenarios.

[0009] To achieve the above purpose, the technical solution adopted by the present invention is as follows:

[0010] A business opportunity capability recommendation system based on collaborative filtering technology, comprising:

[0011] The business opportunity data collection and processing module is used to input real-time business opportunity data from multiple data sources, and clean and pre-process the accessed business opportunity data;

[0012] User behavior data collection and processing module, used to collect user behavior data, clean, preprocess, extract features, normalize the data, and divide it into training sets and test sets;

[0013] Collaborative filtering recommendation engine, which is used to build collaborative filtering models based on business opportunity data and user behavior data, and analyze the relevance and similarity between users and business opportunities;

[0014] The recommendation result generation and display module is used to generate a personalized business opportunity capability recommendation list based on similar user groups calculated by the collaborative filtering model, and display the recommendation results to users;

[0015] Feedback optimization module, which is used to collect user feedback on recommendation results, continuously optimize collaborative filtering models and recommendation algorithms through feedback learning, and improve recommendation accuracy and user satisfaction;

[0016] The data storage and management module is used to store and manage business opportunity data, user behavior data, recommendation results and related information to ensure the data security and stable operation of the system.

[0017] The present invention further illustrates that the collaborative filtering model includes a user-based collaborative filtering recommendation algorithm, a business opportunity-based collaborative filtering recommendation algorithm and a hybrid recommendation algorithm.

[0018] Furthermore, the user-based collaborative filtering recommendation algorithm is to recommend business opportunities to target users by calculating the similarity between users and using the preferences of similar users, including the following steps:

[0019] (1) Similarity calculation;

[0020]

[0021] Among them, r uiis the rating of item i by user u, r vi is the rating of user v on item i, μ u is the mean rating of user u, μ v is the mean rating of user v, I uv is the set of items that user u and user v have jointly evaluated;

[0022] (2) predicting the user's rating of an unrated item;

[0023] For the target user u, the score of the target user u for the unrated item i is predicted by taking the weighted average of the scores of the k users (neighbors) with the highest similarity.

[0024]

[0025] Among them, μ u is the mean rating of user u, μ v is the mean rating of user v, r vi is the rating of user v on item i, N u is the set of k users with the highest similarity to user u, and sim(u,v) is the similarity between user u and user v;

[0026] (3) Based on the predicted ratings, the item with the highest rating is recommended to the user as the recommendation result.

[0027] Furthermore, the business opportunity-based collaborative filtering recommendation algorithm calculates the similarity between business opportunities and uses the user's behavior on similar business opportunities to recommend business opportunities that the user has not interacted with, including the following steps:

[0028] (1) Similarity calculation:

[0029]

[0030] Among them, M al is the rating of user A on business opportunity l, M bl is the score given by user B to business opportunity l, A a is the mean rating of user A, B b is the mean rating of user B, L ab It is the set of business opportunities that user A and user B have evaluated together;

[0031] (2) predicting users’ interest in uninterpreted business opportunities;

[0032] For target user A, the interest level of the target user in the uninteracted business opportunity t is predicted by the ratings of the business opportunities he has interacted with and the similarity of other business opportunities similar to these business opportunities:

[0033] Predicting interest level

[0034] Among them, K a is the set of business opportunities that user A has interacted with, sim(t,k) is the similarity between business opportunities t and k;

[0035] (3) Sort by predicted interest level and recommend the business opportunity with the highest interest to the user as the recommendation result.

[0036] Furthermore, the hybrid recommendation algorithm combines the user-based collaborative filtering recommendation algorithm and the business opportunity-based collaborative filtering recommendation algorithm, linearly combines the recommendation results of the two algorithms according to certain weights, generates a final recommendation list, and transfers it to the feedback optimization module, including the following steps:

[0037] (1) Generate recommendation scores using user-based collaborative filtering recommendation algorithm and business opportunity-based collaborative filtering recommendation algorithm respectively;

[0038] (2) The recommendation score is weighted by the preset weights α and β:

[0039]

[0040] And α+β=1;

[0041] in, is the recommendation score of the user-based collaborative filtering recommendation algorithm, It is the recommendation score of the collaborative filtering recommendation algorithm based on business opportunities;

[0042] (3) Based on the weighted recommendation score, the product with the highest score is recommended to the user as the recommendation result.

[0043] Furthermore, the feedback optimization step of the feedback optimization module includes:

[0044] (1) Regularly use the collected feedback information data as training samples to retrain the collaborative filtering model;

[0045] (2) Before each new recommendation, the system dynamically adjusts the weights of parameters α and β based on the user’s most recent feedback information;

[0046]

[0047] Among them, the weight coefficients α(T) and β(T) are dynamically adjusted with time and user feedback to optimize the recommendation effect;

[0048] (3) Output new recommendations.

[0049] The present invention further illustrates that the business opportunity data collection and processing module collects business opportunity data, such as description, business opportunity type, region, management level, business opportunity source, budget, industry type, etc., from the CRM system, business opportunity system, project system and manual entry through internal data sources, and performs preliminary processing and storage; that is: the business opportunity data collection and processing module uses Apache Kafka to perform real-time data inflow on the above data sources, and at the same time cleans and preprocesses the accessed business opportunity data, including removing duplicate data, processing missing values, and unifying data formats, and finally stores it in the big data storage component Hbase.

[0050] The present invention further illustrates that the user behavior data collection and processing module collects user behavior data (such as website logs, application logs, database queries, API calls, etc.) and cleans the data, including deduplication, filtering, formatting, type conversion, and missing value processing; then extracts features, normalizes data, divides training sets and test sets, and stores the processed data in a suitable database or file system, establishes a data management mechanism, ensures the timeliness and consistency of the data, and provides a high-quality data foundation for the recommendation system.

[0051] The present invention also proposes a business opportunity capability recommendation method based on collaborative filtering technology, which is implemented by the business opportunity capability recommendation system based on collaborative filtering technology, and includes the following steps:

[0052] (1) Collect and pre-process business opportunity data and user behavior data respectively;

[0053] (2) Using user-based collaborative filtering recommendation algorithms to process user behavior data and obtain recommendation results based on user behavior;

[0054] (3) Using the business opportunity-based collaborative filtering recommendation algorithm to process the business opportunity data and obtain recommendation results based on the business opportunity data;

[0055] (4) The recommendation results of step (2) and step (3) are processed using a hybrid recommendation algorithm to obtain the final recommendation result and recommend it to the customer. The final recommendation result is also output to the feedback optimization module, which continuously optimizes the collaborative filtering model and recommendation algorithm through feedback learning to improve the accuracy of the recommendation and user satisfaction.

[0056] The present invention integrates two collaborative filtering algorithms and introduces corresponding hot penalty factors α(T) and β(T) to dynamically adjust the recommendation score, which can effectively reduce the influence of popular users and items on the recommendation and enable long-tail data to receive attention, thereby solving the problems of traditional collaborative filtering algorithms and enhancing the diversity of recommendations.

[0057] By combining user collaborative filtering and business opportunity collaborative filtering, user similarity and business opportunity similarity can be comprehensively utilized to improve the accuracy and diversity of the recommendation system and enhance user satisfaction. Finally, the results are pushed to the feedback optimization module. Assume that the system collects a batch of new business opportunity data through the business opportunity data collection and processing module, and already has historical behavior data of users. The system first analyzes the business opportunity data and user behavior data, builds a collaborative filtering model, and uses the above three algorithm formulas to find other user groups with similar business opportunity preferences. Then, by recommending relevant business opportunity capability resources based on the preferences of these user groups, such as recommending capability providers with similar technical requirements and industry backgrounds. The recommendation results are displayed to users for selection and feedback based on their personalized needs and preferences. The system continuously optimizes the recommendation algorithm based on user feedback information to improve the accuracy and effectiveness of subsequent recommendations.

[0058] Advantages of the present invention:

[0059] 1. Traditional rule-based or content-based recommendation systems are difficult to fully consider the dynamic changes of user preferences and behaviors. The present invention, through collaborative filtering technology, can provide personalized recommendations based on users' historical behaviors and business opportunity data, thereby improving the accuracy of recommendations and user satisfaction.

[0060] 2. The system of the present invention includes a feedback optimization module, which can collect user feedback information on recommendation results and continuously optimize the collaborative filtering model and recommendation algorithm through feedback learning. This dynamic adjustment mechanism ensures that the recommendation system can continuously adapt to changes in user needs, improves the accuracy and effectiveness of recommendations, and makes the recommendation system have better adaptability and scalability, and is suitable for a variety of business environments and application scenarios.

[0061] 3. The collaborative filtering recommendation engine of the system of the present invention includes not only a user-based collaborative filtering algorithm, but also a business opportunity-based collaborative filtering algorithm, as well as a hybrid recommendation algorithm combining the two. This method of comprehensively utilizing user similarity and business opportunity similarity can more comprehensively analyze user behavior and business opportunity characteristics, and improve the accuracy and diversity of recommendations. BRIEF DESCRIPTION OF THE DRAWINGS

[0062] Figure 1 It is a schematic diagram of the system structure of an embodiment of the present invention.

[0063] Figure 2 It is a schematic diagram of a method flow of an embodiment of the present invention. DETAILED DESCRIPTION

[0064] The present invention is further described below in conjunction with the accompanying drawings.

[0065] Example:

[0066] A business opportunity capability recommendation system based on collaborative filtering technology, such as Figure 1 As shown, including:

[0067] The business opportunity data collection and processing module is used to input real-time business opportunity data from multiple data sources, and clean and pre-process the accessed business opportunity data;

[0068] User behavior data collection and processing module, used to collect user behavior data, clean, preprocess, extract features, normalize the data, and divide it into training sets and test sets;

[0069] Collaborative filtering recommendation engine, which is used to build collaborative filtering models based on business opportunity data and user behavior data, and analyze the relevance and similarity between users and business opportunities;

[0070] The recommendation result generation and display module is used to generate a personalized business opportunity capability recommendation list based on similar user groups calculated by the collaborative filtering model, and display the recommendation results to users;

[0071] Feedback optimization module, which is used to collect user feedback on recommendation results, continuously optimize collaborative filtering models and recommendation algorithms through feedback learning, and improve recommendation accuracy and user satisfaction;

[0072] The data storage and management module is used to store and manage business opportunity data, user behavior data, recommendation results and related information to ensure the data security and stable operation of the system.

[0073] It is further explained that in the collaborative filtering recommendation engine, the collaborative filtering model includes a user-based collaborative filtering recommendation algorithm, a business opportunity-based collaborative filtering recommendation algorithm, and a hybrid recommendation algorithm.

[0074] Furthermore, the user-based collaborative filtering recommendation algorithm is to recommend business opportunities to target users by calculating the similarity between users and using the preferences of similar users, including the following steps:

[0075] (1) Similarity calculation;

[0076]

[0077] Among them, r ui is the rating of item i by user u, r vi is the rating of user v on item i, μ u is the mean rating of user u, μ v is the mean rating of user v, I uv is the set of items that user u and user v have jointly evaluated;

[0078] (2) predicting the user's rating of an unrated item;

[0079] For the target user u, the score of the target user u for the unrated item i is predicted by taking the weighted average of the scores of the k users (neighbors) with the highest similarity.

[0080]

[0081] Among them, μ u is the mean rating of user u, μ v is the mean rating of user v, r vi is the rating of user v on item i, N u is the set of k users with the highest similarity to user u, and sim(u,v) is the similarity between user u and user v;

[0082] (3) Based on the predicted ratings, the item with the highest rating is recommended to the user as the recommendation result.

[0083] Furthermore, the business opportunity-based collaborative filtering recommendation algorithm calculates the similarity between business opportunities and uses the user's behavior on similar business opportunities to recommend business opportunities that the user has not interacted with, including the following steps:

[0084] (1) Similarity calculation:

[0085]

[0086] Among them, M al is the rating of user A on business opportunity l, M bl is the score given by user B to business opportunity l, A a is the mean rating of user A, B b is the mean rating of user B, L ab It is the set of business opportunities that user A and user B have evaluated together;

[0087] (2) predicting users’ interest in uninterpreted business opportunities;

[0088] For target user A, the interest level of the target user in the uninteracted business opportunity t is predicted by the ratings of the business opportunities he has interacted with and the similarity of other business opportunities similar to these business opportunities:

[0089]

[0090] Among them, K a is the set of business opportunities that user A has interacted with, sim(t,k) is the similarity between business opportunities t and k;

[0091] (3) Sort by predicted interest level and recommend the business opportunity with the highest interest to the user as the recommendation result.

[0092] Furthermore, the hybrid recommendation algorithm combines the user-based collaborative filtering recommendation algorithm and the business opportunity-based collaborative filtering recommendation algorithm, linearly combines the recommendation results of the two algorithms according to certain weights, generates a final recommendation list, and transfers it to the feedback optimization module, including the following steps:

[0093] (1) Generate recommendation scores using user-based collaborative filtering recommendation algorithm and business opportunity-based collaborative filtering recommendation algorithm respectively;

[0094] (2) The recommendation score is weighted by the preset weights α and β:

[0095]

[0096] And α+β=1;

[0097] in, is the recommendation score of the user-based collaborative filtering recommendation algorithm, It is the recommendation score of the collaborative filtering recommendation algorithm based on business opportunities;

[0098] (3) Based on the weighted recommendation score, the product with the highest score is recommended to the user as the recommendation result.

[0099] Furthermore, the feedback optimization step of the feedback optimization module includes:

[0100] (1) Regularly use the collected feedback information data as training samples to retrain the collaborative filtering model;

[0101] (2) Before each new recommendation, the system dynamically adjusts the weights of parameters α and β based on the user’s most recent feedback information;

[0102]

[0103] Among them, the weight coefficients α(T) and β(T) are dynamically adjusted with time and user feedback to optimize the recommendation effect;

[0104] (3) Output new recommendations.

[0105] It is further explained that the business opportunity data collection and processing module collects business opportunity data, such as description, business opportunity type, region, management level, business opportunity source, budget, industry type, etc., from the CRM system, business opportunity system, project system and manual entry through internal data sources, and performs preliminary processing and storage; that is: the business opportunity data collection and processing module uses Apache Kafka to perform real-time data inflow on the above data sources, and at the same time cleans and preprocesses the accessed business opportunity data, including removing duplicate data, processing missing values, and unifying data formats, and finally stores it in the big data storage component Hbase.

[0106] It is further explained that the user behavior data collection and processing module collects user behavior data (such as website logs, application logs, database queries, API calls, etc.) and cleans the data, including deduplication, filtering, formatting, type conversion, and missing value processing; then extracts features, normalizes data, divides training sets and test sets, and stores the processed data in a suitable database or file system, establishes a data management mechanism, ensures the timeliness and consistency of the data, and provides a high-quality data foundation for the recommendation system.

[0107] This embodiment also proposes a business opportunity capability recommendation method based on collaborative filtering technology, which is implemented by the business opportunity capability recommendation system based on collaborative filtering technology. Figure 2 As shown, the following steps are included:

[0108] (1) Collect and pre-process business opportunity data and user behavior data respectively;

[0109] (2) Using user-based collaborative filtering recommendation algorithms to process user behavior data and obtain recommendation results based on user behavior;

[0110] (3) Using the business opportunity-based collaborative filtering recommendation algorithm to process the business opportunity data and obtain recommendation results based on the business opportunity data;

[0111] (4) The recommendation results of step (2) and step (3) are processed using a hybrid recommendation algorithm to obtain the final recommendation result and recommend it to the customer. The final recommendation result is also output to the feedback optimization module, which continuously optimizes the collaborative filtering model and recommendation algorithm through feedback learning to improve the accuracy of the recommendation and user satisfaction.

[0112] Obviously, the above embodiments are only examples for clearly explaining the present invention, and are not intended to limit the implementation of the present invention. For ordinary technicians in the relevant technical field, other different forms of changes or modifications can be made based on the above description; it is not necessary and impossible to list all the implementation methods here; and the obvious changes or modifications derived from them are still within the protection scope of the present invention.

Claims

1. A business opportunity capability recommendation system based on collaborative filtering technology, characterized in that: include: The business opportunity data collection and processing module is used to input real-time business opportunity data from multiple data sources, and clean and pre-process the accessed business opportunity data; User behavior data collection and processing module, used to collect user behavior data, clean, preprocess, extract features, normalize the data, and divide it into training sets and test sets; Collaborative filtering recommendation engine, which is used to build collaborative filtering models based on business opportunity data and user behavior data, and analyze the relevance and similarity between users and business opportunities; The recommendation result generation and display module is used to generate a personalized business opportunity capability recommendation list based on similar user groups calculated by the collaborative filtering model, and display the recommendation results to users; Feedback optimization module, which is used to collect user feedback on recommendation results, continuously optimize collaborative filtering models and recommendation algorithms through feedback learning, and improve recommendation accuracy and user satisfaction; The data storage and management module is used to store and manage business opportunity data, user behavior data, and recommendation results to ensure the data security and stable operation of the system.

2. The business opportunity capability recommendation system based on collaborative filtering technology according to claim 1, characterized in that: The feedback optimization step of the feedback optimization module includes: (1) Regularly use the collected feedback information data as training samples to retrain the collaborative filtering model; (2) Before each new recommendation, the system dynamically adjusts the weights of parameters α and β based on the user’s most recent feedback information; Among them, the weight coefficients α(T) and β(T) are dynamically adjusted with time and user feedback to optimize the recommendation effect; (3) Output new recommendations.

3. The business opportunity capability recommendation system based on collaborative filtering technology according to claim 1, characterized in that: The collaborative filtering model includes a user-based collaborative filtering recommendation algorithm, a business opportunity-based collaborative filtering recommendation algorithm and a hybrid recommendation algorithm.

4. The business opportunity capability recommendation system based on collaborative filtering technology according to claim 3 is characterized by: The user-based collaborative filtering recommendation algorithm calculates the similarity between users and uses the preferences of similar users to recommend business opportunities to target users, including the following steps: (1) Similarity calculation; Among them, r ui is the rating of item i by user u, r vi is the rating of user v on item i, μ u is the mean rating of user u, μ v is the mean rating of user v, I uv is the set of items that user u and user v have jointly evaluated; (2) predicting the user's rating of an unrated item; For the target user u, the score of the target user u for the unrated item i is predicted by taking the weighted average of the scores of the k users (neighbors) with the highest similarity. Among them, μ u is the mean rating of user u, μ v is the mean rating of user v, r vi is the rating of user v on item i, N u is the set of k users with the highest similarity to user u, and sim(u,v) is the similarity between user u and user v; (3) Based on the predicted ratings, the item with the highest rating is recommended to the user as the recommendation result.

5. The business opportunity capability recommendation system based on collaborative filtering technology according to claim 3 is characterized by: The collaborative filtering recommendation algorithm based on business opportunities calculates the similarity between business opportunities and uses the user's behavior on similar business opportunities to recommend business opportunities that the user has not interacted with, including the following steps: (1) Similarity calculation: Among them, M al is the rating of user A on business opportunity l, M bl is the score given by user B to business opportunity l, A a is the mean rating of user A, B b is the mean rating of user B, L ab It is the set of business opportunities that user A and user B have evaluated together; (2) predicting users’ interest in uninterpreted business opportunities; For target user A, the interest level of the target user in the uninteracted business opportunity t is predicted by the ratings of the business opportunities he has interacted with and the similarity of other business opportunities similar to these business opportunities: Predicting interest level Among them, K a is the set of business opportunities that user A has interacted with, sim(t,k) is the similarity between business opportunities t and k; (3) Sort by predicted interest level and recommend the business opportunity with the highest interest to the user as the recommendation result.

6. The business opportunity capability recommendation system based on collaborative filtering technology according to claim 3 is characterized by: The hybrid recommendation algorithm combines the user-based collaborative filtering recommendation algorithm and the business opportunity-based collaborative filtering recommendation algorithm, linearly combines the recommendation results of the two algorithms according to certain weights, generates a final recommendation list, and transfers it to the feedback optimization module, including the following steps: (1) Generate recommendation scores using user-based collaborative filtering recommendation algorithm and business opportunity-based collaborative filtering recommendation algorithm respectively; (2) The recommendation score is weighted by the preset weights α and β: α+β＝1； in, is the recommendation score of the user-based collaborative filtering recommendation algorithm, It is the recommendation score of the collaborative filtering recommendation algorithm based on business opportunities; (3) Based on the weighted recommendation score, the product with the highest score is recommended to the user as the recommendation result.

7. The business opportunity capability recommendation system based on collaborative filtering technology according to claim 1, characterized in that: The business opportunity data collection and processing module collects business opportunity data from the CRM system, business opportunity system, project system and manual entry through internal data sources, and performs preliminary processing and storage; that is: the business opportunity data collection and processing module uses Apache Kafka to perform real-time data inflow on the above data sources, and at the same time cleans and preprocesses the accessed business opportunity data, including removing duplicate data, processing missing values, and unifying data formats, and finally storing it in the big data storage component Hbase.

8. The business opportunity capability recommendation system based on collaborative filtering technology according to claim 1, characterized in that: The user behavior data collection and processing module collects user behavior data and cleans the data, including deduplication, filtering, formatting, type conversion, and missing value processing; then extracts features, normalizes data, divides training sets and test sets, and stores the processed data in a suitable database or file system, establishes a data management mechanism, ensures the timeliness and consistency of the data, and provides a high-quality data foundation for the recommendation system.

9. A business opportunity capability recommendation method based on collaborative filtering technology, characterized by: The method is implemented by a business opportunity capability recommendation system based on collaborative filtering technology as claimed in any one of claims 1 to 8, comprising the following steps: (1) Collect and pre-process business opportunity data and user behavior data respectively; (2) Using user-based collaborative filtering recommendation algorithms to process user behavior data and obtain recommendation results based on user behavior; (3) Using the business opportunity-based collaborative filtering recommendation algorithm to process the business opportunity data and obtain recommendation results based on the business opportunity data; (4) The recommendation results of step (2) and step (3) are processed using a hybrid recommendation algorithm to obtain the final recommendation result and recommend it to the customer. The final recommendation result is also output to the feedback optimization module, which continuously optimizes the collaborative filtering model and recommendation algorithm through feedback learning to improve the accuracy of the recommendation and user satisfaction.

Images