Intelligent recommendation method and device for campus micro-service, recommendation equipment and storage medium

Abstract

The invention relates to an intelligent recommendation method and device for campus micro-services, recommendation equipment and a storage medium, and belongs to the technical field of data personalized recommendation, the intelligent recommendation method for the campus micro-services comprises the following steps: obtaining metadata of a plurality of micro-services, performing vectorization coding on the preprocessed metadata to obtain a plurality of feature vectors, and storing the feature vectors in a database; constructing a micro-service feature matrix based on the plurality of feature vectors; obtaining an initial query statement of the target user, and generating a plurality of query vectors based on the initial query statement, the historical behavior of the target user and the service association rule base; and determining a candidate feature vector set in the micro-service feature matrix based on the plurality of query vectors, and determining a micro-service recommendation list corresponding to the target user based on the historical behaviors of the target user, the historical behaviors of the plurality of users and the candidate feature vector set. According to the invention, the efficiency of micro-service intelligent recommendation is effectively improved, and intelligent service guidance of campus micro-services is realized.

Description

Technical Field

[0001] This invention relates to the field of personalized data recommendation technology, and in particular to an intelligent recommendation method, apparatus, recommendation device, and storage medium for campus microservices. Background Technology

[0002] The campus microservice system is one of the core carriers of campus digital services. When teachers and students need to use a specific service, the traditional access method requires users to search through the portal page layer by layer or remember the specific application entry, which has problems such as difficulty in service discovery and lengthy operation paths.

[0003] To improve service efficiency, the service access experience can currently be optimized through application categorization and search functions. However, due to the diversity of campus services and the continuous growth in the number of applications, actual research has found that faculty and student users spend most of their time repeatedly browsing and searching for the target service entry point, and frequently encounter situations where they "cannot find" or "are unaware that the relevant service exists."

[0004] Currently, most universities lack intelligent service guidance mechanisms when providing digital services. They typically use static portal directories or basic search functions for faculty and students to find applications, requiring users to navigate through levels or memorize keywords to locate target services. These solutions only provide basic service navigation, resulting in unresolved issues such as low service discovery efficiency, lengthy service access paths, and the invisibility of hidden services.

[0005] Therefore, how to achieve intelligent service guidance in campus microservice systems has become an urgent technical problem to be solved. Summary of the Invention

[0006] In view of this, it is necessary to provide an intelligent recommendation method, device, recommendation equipment, and storage medium for campus microservices, so as to realize intelligent service guidance in the campus microservice system.

[0007] To achieve the above objectives, in a first aspect, the present invention provides an intelligent recommendation method for campus microservices, comprising: Obtain metadata from multiple microservices, vectorize the preprocessed metadata to obtain multiple feature vectors, and construct a microservice feature matrix based on the multiple feature vectors; Obtain the initial query statement of the target user, and generate multiple query vectors based on the initial query statement, the target user's historical behavior, and the service association rule base. The service association rule base is constructed based on the historical behavior of multiple users. Based on multiple query vectors, a candidate feature vector set is determined in the microservice feature matrix. Then, based on the target user's historical behavior, the historical behavior of multiple users, and the candidate feature vector set, a microservice recommendation list corresponding to the target user is determined.

[0008] In one possible implementation, the metadata of the multiple microservices includes: name text, description text, tag text, and permission text.

[0009] In one possible implementation, the vectorization encoding of the preprocessed metadata yields multiple feature vectors, including: The preprocessed metadata is vectorized and encoded using the Embedding model to obtain multiple feature vectors.

[0010] In one possible implementation, the generation of multiple query vectors based on the initial query statement, the target user's historical behavior, and a service association rule base includes: Vectorize the initial query statement to obtain the first query vector; The initial query statement is expanded based on the service association rule base to obtain an expanded query statement, and the expanded query statement is vectorized to obtain a second query vector; The related keywords are extracted from the target user's historical behavior and added to the initial query statement to obtain the related query statement. The related query statement is then vectorized to obtain the third query vector.

[0011] In one possible implementation, determining the candidate feature vector set based on multiple query vectors in the microservice feature matrix includes: A predetermined number of feature vectors with the highest cosine similarity to multiple query vectors in the microservice feature matrix are selected as the candidate feature vector set.

[0012] In one possible implementation, determining the microservice recommendation list corresponding to the target user based on the target user's historical behavior, the historical behavior of multiple users, and a candidate feature vector set includes: The target microservice set is determined based on the target user's historical behavior and the historical behavior of multiple users. The target microservice set includes microservices that are associated with the target user's historical behavior, that the target user has not used, and that have been used by other users more than a usage threshold. A list of recommended microservices for a target user is determined based on the target set of microservices and the candidate feature vector set.

[0013] In one possible implementation, determining the microservice recommendation list corresponding to the target user based on the target microservice set and the candidate feature vector set includes: Based on the candidate feature vector set, a candidate microservice set is determined. Duplicate microservices in the target microservice set and the candidate microservice set are removed. The remaining microservices in the target microservice set and the candidate microservice set are determined as the microservice recommendation list corresponding to the target user.

[0014] On the other hand, the present invention also provides an intelligent recommendation device for campus microservices, comprising: The acquisition module is used to acquire metadata from multiple microservices, vectorize the preprocessed metadata to obtain multiple feature vectors, and construct a microservice feature matrix based on the multiple feature vectors. The generation module is used to obtain the initial query statement of the target user, and generate multiple query vectors based on the initial query statement, the target user's historical behavior, and the service association rule base. The service association rule base is constructed based on the historical behavior of multiple users. The determination module is used to determine the candidate feature vector set in the microservice feature matrix based on multiple query vectors, and to determine the microservice recommendation list corresponding to the target user based on the target user's historical behavior, the historical behavior of multiple users, and the candidate feature vector set.

[0015] Secondly, the present invention also provides a recommended device, comprising a memory and a processor, wherein, The memory is used to store programs; The processor, coupled to the memory, is used to execute the program stored in the memory to implement the steps in the intelligent recommendation method for campus microservices described in any of the above implementations.

[0016] Thirdly, the present invention also provides a computer-readable storage medium for storing a computer-readable program or instruction, which, when executed by a processor, can implement the steps of the intelligent recommendation method for campus microservices described in any of the above implementations.

[0017] The beneficial effects of this invention are as follows: The intelligent recommendation method, apparatus, recommendation device, and storage medium for campus microservices provided by this invention construct a microservice feature matrix through the metadata of the microservices, providing a data foundation for intelligent recommendation of microservices. Then, multiple query vectors are generated through the initial query statement of the target user, providing direction for intelligent recommendation of microservices. This determines a set of candidate feature vectors in the microservice feature matrix. Finally, the candidate feature vector set is expanded by combining the historical behavior of the target user and the historical behavior of multiple users to determine the microservice recommendation list corresponding to the target user. This invention effectively improves the efficiency of intelligent recommendation of microservices and realizes intelligent service guidance for campus microservices. Attached Figure Description

[0018] Figure 1 A schematic flowchart of an embodiment of the intelligent recommendation method for campus microservices provided by the present invention; Figure 2 A schematic flowchart of an embodiment of the intelligent recommendation process for campus microservices provided by the present invention; Figure 3A schematic diagram of an embodiment of the intelligent recommendation device for campus microservices provided by the present invention; Figure 4 A schematic diagram of an embodiment of the recommended device provided by the present invention. Detailed Implementation

[0019] 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 a part of the embodiments of the present invention, and not all of them. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0020] In the description of the embodiments of the present invention, unless otherwise stated, "multiple" means two or more. "And / or" describes the relationship between related objects, indicating that there can be three relationships. For example, A and / or B can represent three situations: A exists alone, A and B exist simultaneously, and B exists alone.

[0021] The terms "first," "second," etc., used in the embodiments of this invention are for descriptive purposes only and should not be construed as indicating or implying their relative importance or implicitly specifying the number of technical features indicated. Therefore, a technical feature defined with "first" or "second" may explicitly or implicitly include at least one of that feature.

[0022] In this document, the term "embodiment" means that a particular feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of the invention. The appearance of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment mutually exclusive with other embodiments. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.

[0023] This invention provides an intelligent recommendation method, apparatus, recommendation device, and storage medium for campus microservices, which will be described below.

[0024] Figure 1 A schematic flowchart of an embodiment of the intelligent recommendation method for campus microservices provided by the present invention is shown below. Figure 1 As shown, the intelligent recommendation methods for campus microservices include: S101. Obtain metadata from multiple microservices, perform vectorization encoding on the preprocessed metadata to obtain multiple feature vectors, and construct a microservice feature matrix based on the multiple feature vectors.

[0025] It should be noted that the intelligent recommendation method for campus microservices provided by this invention can be applied to data-driven personalized recommendation scenarios, especially intelligent recommendation scenarios for campus microservices.

[0026] When making intelligent recommendations for campus microservices, the recommendation device (such as a laptop, desktop computer, or mobile phone) can first obtain the metadata of multiple microservices, then preprocess the metadata (such as removing duplicate and invalid data), then vectorize the preprocessed metadata to obtain multiple feature vectors, and then construct a microservice feature matrix based on the multiple feature vectors to provide a data foundation for intelligent recommendations of microservices.

[0027] S102. Obtain the initial query statement of the target user, generate multiple query vectors based on the initial query statement, the target user's historical behavior, and the service association rule base, and construct the service association rule base based on the historical behavior of multiple users.

[0028] It's important to note that after constructing the microservice feature matrix, the initial query statement of the target user can be obtained to initially assess their query preferences. Then, multiple query vectors can be generated using the initial query statement, the target user's historical behavior, and the service association rule base to improve the hit rate of the recommendation service. The service association rule base, constructed based on the historical behavior of multiple users, can be used to indicate the query terms associated with the initial query statement.

[0029] S103. Based on multiple query vectors, determine the candidate feature vector set in the microservice feature matrix, and based on the target user's historical behavior, the historical behavior of multiple users, and the candidate feature vector set, determine the microservice recommendation list corresponding to the target user.

[0030] It should be noted that after determining multiple query vectors, a candidate feature vector set can be determined in the microservice feature matrix based on the multiple query vectors. Then, the target user's historical behavior, the historical behavior of multiple users, and the candidate feature vector set can be combined to determine the microservice recommendation list corresponding to the target user, thereby realizing intelligent recommendation of microservices.

[0031] In summary, the intelligent recommendation method for campus microservices provided in this invention constructs a microservice feature matrix through the metadata of the microservices, providing a data foundation for intelligent recommendation. Then, it generates multiple query vectors based on the initial query statement of the target user, providing direction for intelligent recommendation. This determines a set of candidate feature vectors in the microservice feature matrix. Finally, it expands the candidate feature vector set by combining the historical behavior of the target user with the historical behavior of multiple users, determining the microservice recommendation list corresponding to the target user. This invention effectively improves the efficiency of intelligent recommendation of microservices and realizes intelligent service guidance for campus microservices.

[0032] In some embodiments of the present invention, the metadata of the plurality of microservices may include: name text, description text, tag text, and permission text.

[0033] In some embodiments of the present invention, the step of vectorizing the preprocessed metadata to obtain multiple feature vectors includes: The preprocessed metadata is vectorized and encoded using the Embedding model to obtain multiple feature vectors.

[0034] It should be noted that when vectorizing the preprocessed metadata to obtain multiple feature vectors, the embedding model can be used to vectorize the preprocessed metadata to obtain multiple feature vectors. The subsequent vectorization process in this invention can also be implemented using the embedding model.

[0035] In some embodiments of the present invention, the generation of multiple query vectors based on the initial query statement, the target user's historical behavior, and the service association rule base includes: Vectorize the initial query statement to obtain the first query vector; The initial query statement is expanded based on the service association rule base to obtain an expanded query statement, and the expanded query statement is vectorized to obtain a second query vector; The related keywords are extracted from the target user's historical behavior and added to the initial query statement to obtain the related query statement. The related query statement is then vectorized to obtain the third query vector.

[0036] It's important to note that when generating multiple query vectors based on the initial query, the target user's historical behavior, and the service association rule base, the initial query can be directly vectorized to obtain the first query vector. Then, the initial query is expanded using the service association rule base to obtain an expanded query, which is then vectorized to obtain the second query vector. Next, related terms can be extracted from the target user's historical behavior and added to the initial query to obtain related query statements, which are then vectorized to obtain the third query vector. By generating multiple query vectors, multi-dimensional requirements such as query semantics, keywords, and user behavior can be covered, improving the hit rate of the target service.

[0037] In some embodiments of the present invention, determining a candidate feature vector set based on multiple query vectors in a microservice feature matrix includes: A predetermined number of feature vectors with the highest cosine similarity to multiple query vectors in the microservice feature matrix are selected as the candidate feature vector set.

[0038] It should be noted that when determining the candidate feature vector set in the microservice feature matrix based on multiple query vectors, a preset number of feature vectors with the highest cosine similarity to the multiple query vectors in the microservice feature matrix can be determined as the candidate feature vector set.

[0039] In some embodiments of the present invention, determining the microservice recommendation list corresponding to the target user based on the target user's historical behavior, the historical behavior of multiple users, and a candidate feature vector set includes: The target microservice set is determined based on the target user's historical behavior and the historical behavior of multiple users. The target microservice set includes microservices that are associated with the target user's historical behavior, that the target user has not used, and that have been used by other users more than a usage threshold. A list of recommended microservices for a target user is determined based on the target set of microservices and the candidate feature vector set.

[0040] It should be noted that when determining the microservice recommendation list corresponding to the target user based on the target user's historical behavior, the historical behavior of multiple users, and the candidate feature vector set, a target microservice set can be determined based on the target user's historical behavior and the historical behavior of multiple users. This target microservice set includes microservices associated with the target user's historical behavior that the target user has not used and that have been used by other users more than a usage threshold. For example, a large language model can be trained using the historical behavior of multiple users, and then this model can be used to determine the target microservice set based on the target user's historical behavior and the historical behavior of multiple users. Then, the microservice recommendation list corresponding to the target user can be determined based on the target microservice set and the candidate feature vector set.

[0041] In some embodiments of the present invention, determining the microservice recommendation list corresponding to the target user based on the target microservice set and the candidate feature vector set includes: Based on the candidate feature vector set, a candidate microservice set is determined. Duplicate microservices in the target microservice set and the candidate microservice set are removed. The remaining microservices in the target microservice set and the candidate microservice set are determined as the microservice recommendation list corresponding to the target user.

[0042] It should be noted that when determining the microservice recommendation list corresponding to the target user based on the target microservice set and the candidate feature vector set, duplicate microservices in the target microservice set and the candidate microservice set can be removed, and the remaining microservices can be determined as the microservice recommendation list corresponding to the target user.

[0043] Combination Figure 2 The microservice intelligent recommendation process provided by this invention specifically includes the following steps: 1. Microservice metadata collection and preprocessing.

[0044] collection The structured metadata for each microservice item consists of text in four dimensions: "name, description, tags, and permissions".

[0045] Preprocessing steps such as filtering, processing, and enhancing the raw metadata text yield microservice metadata: .

[0046] 2. Feature encoding and weighted fusion.

[0047] The Embedding model is used to vectorize and encode the four-dimensional metadata text of the microservices, generate feature vectors, and construct the feature matrix of microservice items.

[0048] 3. User intent parsing and vectorization.

[0049] The overall strategy employs a multi-path query generation and hybrid recall approach: Original Query: Directly vectorized.

[0050] Rewrite Query: Generate extended queries based on the semantics of the original query according to the predefined service association rule base, for example: "grades" is associated with "GPA query + exam schedule".

[0051] Keyword expansion query: Extract related words from user history behavior. For example, if a user frequently visits the "library", add the keyword "reservation".

[0052] An encoder (Embedding model) is used to generate feature vectors for three different query variants, which are then used for subsequent recall and ranking. This multi-path query generation and hybrid recall avoids the failure of a single strategy, covers multiple dimensions of query semantics, keywords, and user behavior, and improves the hit rate of the target service.

[0053] 4. Define the service association rule base.

[0054] 1) Requirements analysis.

[0055] Define your objectives: First, determine the specific objectives of the expanded functionality. For example, do you want the expanded results to more comprehensively cover multiple aspects related to the original question that users might be interested in, or do you want to guide users to use certain specific functional modules? For instance, for the input "grades," if the goal is to help students understand grade-related matters more comprehensively, then you need to consider multiple aspects such as grade inquiries, exam arrangements, and grade analysis.

[0056] User group analysis: Analyze the characteristics of the user group using the application. If it is a student group, then it may be necessary to consider the lexical associations related to learning, exams, etc.; if it is a faculty and staff group, it may involve associations with performance, training, etc. Understand the usage habits and common needs of the user group. For example, students may be more accustomed to using the term "GPA", while in some schools, concepts such as "average score" may be more commonly used. These all need to be considered when building the rule base.

[0057] 2) Data collection and collation.

[0058] Collect corpus: Collect a large amount of text data related to the application field. This data can come from users' historical query records, documents in the professional field, forum discussions, etc. For example, for an application in the education field, various expressions regarding grades, courses, exams, etc. can be collected from school official websites, education forums, student communication groups, etc.

[0059] Data preprocessing: Clean the collected text data to remove irrelevant noise data, such as advertisements, meaningless symbols, etc. Perform word segmentation on the text to split sentences into individual lexical or phrase units. For example, split the sentence "The grade query system has been updated" into words such as "grade", "query", "system", "has", "been", "updated", etc.词性标注 can also be carried out to understand the grammatical functions of each word in the sentence, which is helpful for judging the semantic associations between words when building the rule base later.

[0060] 3) Association rule mining.

[0061] Use statistical methods to mine the associations between words. For example, calculate the frequency of co-occurrence of word pairs (such as "grade" and "query") in the text. If they often appear together, it can be considered that there is an association between them.

[0062] Calculate indicators such as confidence and support. Support represents the proportion of co-occurrence of word pairs in all text data, and confidence represents the probability that the second word appears in the text containing the first word. By setting certain thresholds, select the qualified associated word pairs. For example, if the frequency of co-occurrence of "grade" and "query" is very high in a large amount of text and the confidence is also relatively high, they can be regarded as an association rule.

[0063] Use semantic analysis tools in natural language processing technology. For example, through word embedding technology, map words into a high-dimensional space, where semantically similar words will cluster together. For the input word, its neighboring words in the semantic space can be found as associated words. For example, "grade" may be close to words such as "GPA", "score", "rank" in the semantic space.

[0064] By combining knowledge graphs, which contain rich semantic relationships, such as "grades" being an attribute of "students' academic performance" and "exam arrangements" being related to "students' academic performance", these semantic connections can be mined through path lookup and other methods in knowledge graphs.

[0065] 4) Rule base construction and optimization.

[0066] Building a rule base: Organize the discovered association rules into a rule base. The rule base can be in the form of key-value pairs, where the key is the word entered by the user, and the value is the set of words associated with it.

[0067] For example, the key "grades" could have values ​​such as "GPA lookup", "exam schedule", and "grade analysis".

[0068] The rule base is stored in a database for easy querying and updating.

[0069] Rule optimization: The rule base is manually reviewed and adjusted. This is because automatically generated rules may contain noise, such as unreasonable related words.

[0070] Optimize based on user feedback. For example, if users report that the expanded version of a certain input word does not meet their expectations, the relevant rules need to be adjusted.

[0071] The rule base should be updated regularly, continuously adding new association rules and eliminating outdated rules as application areas develop and user needs change. For example, with the implementation of a new examination system, it may be necessary to add association rules for words related to the new system.

[0072] 5. Extraction of historical behavior data.

[0073] Data Collection: Collect various user behavior data within the application, including but not limited to clicks, searches, browsing history, and dwell time. For example, record the frequency and timing of user visits to library-related pages. This data can be recorded through the application's in-app logging system to ensure data integrity and accuracy.

[0074] User profile building: Based on the collected behavioral data, a profile is built for each user. The profile includes information such as the user's interests, frequently used functions, and behavioral patterns.

[0075] For example, if a user frequently visits the library page, then "library" can be tagged as one of the user's interests in the user profile.

[0076] For each user profile's points of interest, analyze other related behavioral data to uncover potential related keywords.

[0077] For example, for the interest point "library", analyze the user's specific behavior on the library page, such as whether they clicked the reservation button or checked the library's opening hours.

[0078] If a user clicks the reservation button multiple times, then it can be assumed that "reservation" is associated with "library," and "reservation" can be used as a related keyword for "library."

[0079] Statistical methods are used to calculate the correlation between different words or functions in user behavior. For example, the frequency with which users perform a "reservation" action when visiting the "library" page can be calculated.

[0080] Pattern recognition technology was used to identify common patterns in user behavior. For example, it was found that many users would make a reservation after visiting the library page, indicating that "reservation" is a behavior strongly associated with "library".

[0081] Based on the identified related terms, keyword expansion rules are generated. These rules define which related terms should be added under what circumstances.

[0082] For example, a rule could be defined as: "If a user visits the library page more than 5 times in the past month and makes a reservation, add 'reservation' to the keyword expansion."

[0083] The generated rules are stored in a rule base for easy querying and application. The rule base is stored using a database.

[0084] The rules in the rule base consist of two parts: conditions and results. The conditions are the characteristics of user behavior, and the results are the related terms that need to be added.

[0085] When a user performs a search or query, the system expands the user's keywords in real time based on the user profile and rules in the rule base.

[0086] For example, if a user enters "library", the system will find that the user has a strong association with "reservation" based on the records in the user profile and the rules in the rule base, so the search results will automatically expand to "library reservation".

[0087] Based on the expanded keywords, the system can provide users with more personalized search results or recommendations. For example, when a user searches for "library reservation," the system can recommend relevant content such as the library's reservation system portal and reservation process instructions.

[0088] 6. Similarity matching and ranking.

[0089] Indexing the vector library using a large model, and calculating... The cosine similarity with each vector in the feature matrix vector library, for each The four-dimensional similarity vector Perform a weighted calculation to obtain Corresponding cosine similarity Finally obtained and Similarity vector The algorithm outputs a TopK list of microservices and sorts them in descending order of similarity. Tags are combined with user behavior data such as "likes" and "dislikes," and preference dependencies are implemented through weights. A coarse candidate set is generated from all microservices using vector similarity, and then refined by incorporating user behavior feedback weights to generate the final recommendation list.

[0090] In addition, intelligent application recommendations can be achieved through collaborative filtering (CF) algorithms based on user behavior and system data.

[0091] 1. Microservice global feature modeling.

[0092] Construct a global feature matrix for microservices, including four dimensions: global clicks, global favorites, clicks in the last 30 days, and favorites in the last 30 days.

[0093] Constructing application feature moments First, normalization is performed, then collaborative filtering or matrix factorization is used to obtain the user matrix and application matrix, then matrix dimensionality reduction is performed and similarity is calculated to obtain the recommendation service.

[0094] 2. Modeling of personalized user behavior characteristics.

[0095] Construct a user behavior data feature vector. For each microservice item, record the number of clicks, favorites, and dwell time of the user on that microservice item. Build a user index model and adopt a hybrid recommendation strategy.

[0096] Collaborative filtering approach: Based on clustering algorithms, similar user groups are found, and frequently used services within the groups are recommended.

[0097] Matrix factorization path: Matrix factorization yields a user matrix and an application matrix, followed by matrix dimensionality reduction and similarity calculation to obtain the recommendation service.

[0098] The two methods are combined to generate the final microservice recommendation list.

[0099] 3. Recommendation based on a large model. Key steps are as follows: Feature concatenation: Concatenates the global matrix and individual vectors into structured text. For example: "User A clicked on {grade inquiry, library reservation} and bookmarked {academic affairs system} in the past 30 days; popular services across the university: course selection (1204 clicks) and repair request (918 clicks)."

[0100] The system extracts the service name using regular expressions, provides the service name and link, and allows users to directly access the service page by clicking the link in the reply. It transforms the structured feature matrix into a natural language description, leverages the complex reasoning capabilities of a large language model to achieve cross-domain recommendations, and solves the data sparsity problem of collaborative filtering.

[0101] 4. User interest and campus-wide trend service recommendation process.

[0102] User behavior data collection: Collect user behavior data in the application, including clicks, favorites, usage frequency, etc.

[0103] For example, it records that user A clicked on "Grade Inquiry" and "Library Reservation" in the past 30 days, and bookmarked "Academic Affairs System".

[0104] Global data collection: Collect service usage data across the entire campus, including click volume and usage frequency of each service.

[0105] For example, the most popular services across the university are "course selection" (1204 clicks) and "repair request" (918 clicks).

[0106] Feature concatenation: Concatenates user behavior data and global data into structured text.

[0107] Example: User A clicked on {Grade Inquiry, Library Reservation} and bookmarked {Academic Affairs System} in the past 30 days; Popular services across the university: Course Selection (1204 clicks) and Repair Request (918 clicks).

[0108] Prompt word engineering: Design a clear prompt word template to guide the large model in generating recommendation results.

[0109] Fill in the prompt words: Fill the prompt word template with the concatenated user characteristic text.

[0110] Recommendation logic: User interest analysis: Analyze user behavior data to extract users' main interests.

[0111] For example, user A's interest in "grade inquiry" and "library reservation" may indicate a greater focus on academic and campus life services.

[0112] Campus-wide trend analysis: Analyze popular services across the campus to understand the services that are currently most popular on campus.

[0113] For example, the most popular services across the university are "course selection" and "repair requests," indicating that these services are in high demand among students.

[0114] Recommendation generation: A recommendation list is generated by combining user interests and trends across the entire school.

[0115] Ensure that the recommended services are ones that the user has not used before, in order to avoid duplicate recommendations.

[0116] Example recommendation logic: User interest expansion: Based on the services a user has used, recommend other related services. For example, if a user uses "grade inquiry," you can recommend "GPA analysis" or "exam scheduling."

[0117] Campus-wide trend guidance: Recommend popular services across the campus that users haven't used. For example, the popular "course selection" service can be recommended if users haven't used it.

[0118] Overall Recommendation: Combining user interests and campus-wide trends, we recommend services that align with both user interests and campus-wide trends.

[0119] Output recommendation results: Output recommendation results in the specified format.

[0120] Traditional solutions fail to effectively improve the efficiency of students and faculty accessing target applications in campus digital services, often requiring users to spend significant time browsing or searching to locate the services they need. This solution, through an intelligent recommendation engine and natural language interaction technology, automatically analyzes user behavior characteristics and query intent, accurately matching and proactively recommending relevant service entry points. This significantly improves service discovery efficiency, optimizes service access paths, and increases the utilization rate of implicit services. This solution addresses the core pain points of traditional campus portals—"difficult to find services and lengthy paths"—leading to a qualitative leap in the efficiency of digital service access.

[0121] To better implement the intelligent recommendation method for campus microservices in this invention, based on the intelligent recommendation method for campus microservices, the corresponding method is as follows: Figure 3 As shown, this embodiment of the invention also provides an intelligent recommendation device for campus microservices. The intelligent recommendation device 300 for campus microservices includes: The acquisition module 301 is used to acquire metadata of multiple microservices, perform vectorization encoding on the preprocessed metadata to obtain multiple feature vectors, and construct a microservice feature matrix based on the multiple feature vectors. The generation module 302 is used to obtain the initial query statement of the target user, and generate multiple query vectors based on the initial query statement, the target user's historical behavior, and the service association rule base. The service association rule base is constructed based on the historical behavior of multiple users. The determination module 303 is used to determine a candidate feature vector set in the microservice feature matrix based on multiple query vectors, and to determine the microservice recommendation list corresponding to the target user based on the target user's historical behavior, the historical behavior of multiple users, and the candidate feature vector set.

[0122] The intelligent recommendation device 300 for campus microservices provided in the above embodiments can implement the technical solutions described in the above embodiments of the intelligent recommendation method for campus microservices. The specific implementation principles of each module or unit can be found in the corresponding content in the above embodiments of the intelligent recommendation method for campus microservices, and will not be repeated here.

[0123] like Figure 4 As shown, the present invention also provides a recommended device 400. The recommended device 400 includes a processor 401, a memory 402, and a display 403. Figure 4 Only some of the components of the recommended device 400 are shown, but it should be understood that it is not required to implement all of the components shown, and more or fewer components may be implemented instead.

[0124] In some embodiments, processor 401 may be a central processing unit (CPU), microprocessor, or other data processing chip, used to run program code stored in memory 402 or process data, such as the intelligent recommendation method for campus microservices in this invention.

[0125] In some embodiments, processor 401 may be a single server or a group of servers. The server group may be centralized or distributed. In some embodiments, processor 401 may be local or remote. In some embodiments, processor 401 may be implemented on a cloud platform. In one embodiment, the cloud platform may include a private cloud, public cloud, hybrid cloud, community cloud, distributed cloud, internal cloud, multi-cloud, etc., or any combination thereof.

[0126] In some embodiments, memory 402 may be an internal storage unit of the recommended device 400, such as a hard disk or memory of the recommended device 400. In other embodiments, memory 402 may also be an external storage device of the recommended device 400, such as a pluggable hard disk, smart media card (SMC), secure digital (SD) card, flash card, etc., provided on the recommended device 400.

[0127] Furthermore, the memory 402 may include both internal storage units of the recommended device 400 and external storage devices. The memory 402 is used to store application software and various types of data installed on the recommended device 400.

[0128] In some embodiments, display 403 may be an LED display, a liquid crystal display, a touch-sensitive liquid crystal display, or an organic light-emitting diode (OLED) touchscreen. Display 403 is used to display information from the recommending device 400 and to display a visual user interface. Components 401-403 of the recommending device 400 communicate with each other via a system bus.

[0129] In one embodiment, when processor 401 executes the intelligent recommendation program for campus microservices in memory 402, the following steps can be implemented: Obtain metadata from multiple microservices, vectorize the preprocessed metadata to obtain multiple feature vectors, and construct a microservice feature matrix based on the multiple feature vectors; Obtain the initial query statement of the target user, and generate multiple query vectors based on the initial query statement, the target user's historical behavior, and the service association rule base. The service association rule base is constructed based on the historical behavior of multiple users. Based on multiple query vectors, a candidate feature vector set is determined in the microservice feature matrix. Then, based on the target user's historical behavior, the historical behavior of multiple users, and the candidate feature vector set, a microservice recommendation list corresponding to the target user is determined.

[0130] It should be understood that when the processor 401 executes the intelligent recommendation program for the campus microservices in the memory 402, in addition to the functions mentioned above, it can also perform other functions, as detailed in the description of the corresponding method embodiments above.

[0131] Furthermore, this embodiment of the invention does not specifically limit the type of the recommended device 400 mentioned. The recommended device 400 can be a mobile phone, tablet computer, personal digital assistant (PDA), wearable device, laptop computer, or other portable electronic device. Exemplary embodiments of portable electronic devices include, but are not limited to, portable electronic devices running iOS, Android, Microsoft, or other operating systems. The aforementioned portable electronic devices can also be other portable electronic devices, such as laptop computers with touch-sensitive surfaces (e.g., touch panels). It should also be understood that in some other embodiments of the invention, the recommended device 400 may not be a portable electronic device, but rather a desktop computer with a touch-sensitive surface (e.g., a touch panel).

[0132] Accordingly, this application also provides a computer-readable storage medium for storing computer-readable programs or instructions. When the programs or instructions are executed by a processor, they can implement the steps or functions of the intelligent recommendation method for campus microservices provided in the above-described method embodiments.

[0133] Those skilled in the art will understand that all or part of the processes of the methods described in the above embodiments can be implemented by a computer program instructing related hardware (such as a processor, controller, etc.), and the computer program can be stored in a computer-readable storage medium. The computer-readable storage medium may be a disk, optical disk, read-only memory, or random access memory, etc.

[0134] The above provides a detailed description of the intelligent recommendation method, apparatus, recommendation device, and storage medium for campus microservices provided by this invention. Specific examples have been used to illustrate the principles and implementation methods of this invention. The descriptions of the above embodiments are only for the purpose of helping to understand the method and core ideas of this invention. At the same time, those skilled in the art will recognize that there will be changes in the specific implementation methods and application scope based on the ideas of this invention. Therefore, the content of this specification should not be construed as a limitation of this invention.

Claims

1. An intelligent recommendation method for campus microservices, characterized in that, include: Obtain metadata from multiple microservices, vectorize the preprocessed metadata to obtain multiple feature vectors, and construct a microservice feature matrix based on the multiple feature vectors; Obtain the initial query statement of the target user, and generate multiple query vectors based on the initial query statement, the target user's historical behavior, and the service association rule base. The service association rule base is constructed based on the historical behavior of multiple users. Based on multiple query vectors, a candidate feature vector set is determined in the microservice feature matrix. Then, based on the target user's historical behavior, the historical behavior of multiple users, and the candidate feature vector set, a microservice recommendation list corresponding to the target user is determined.

2. The intelligent recommendation method for campus microservices according to claim 1, characterized in that, The metadata of the multiple microservices includes: name text, description text, tag text, and permission text.

3. The intelligent recommendation method for campus microservices according to claim 2, characterized in that, The preprocessed metadata is vectorized and encoded to obtain multiple feature vectors, including: The preprocessed metadata is vectorized and encoded using the Embedding model to obtain multiple feature vectors.

4. The intelligent recommendation method for campus microservices according to claim 1, characterized in that, The process generates multiple query vectors based on the initial query statement, the target user's historical behavior, and the service association rule base, including: Vectorize the initial query statement to obtain the first query vector; The initial query statement is expanded based on the service association rule base to obtain an expanded query statement, and the expanded query statement is vectorized to obtain a second query vector; The related keywords are extracted from the target user's historical behavior and added to the initial query statement to obtain the related query statement. The related query statement is then vectorized to obtain the third query vector.

5. The intelligent recommendation method for campus microservices according to claim 1, characterized in that, The process of determining a candidate feature vector set based on multiple query vectors in the microservice feature matrix includes: A predetermined number of feature vectors with the highest cosine similarity to multiple query vectors in the microservice feature matrix are selected as the candidate feature vector set.

6. The intelligent recommendation method for campus microservices according to claim 1, characterized in that, The process of determining the microservice recommendation list corresponding to the target user based on the target user's historical behavior, the historical behavior of multiple users, and candidate feature vector sets includes: The target microservice set is determined based on the target user's historical behavior and the historical behavior of multiple users. The target microservice set includes microservices that are associated with the target user's historical behavior, that the target user has not used, and that have been used by other users more than a usage threshold. A list of recommended microservices for a target user is determined based on the target set of microservices and the candidate feature vector set.

7. The intelligent recommendation method for campus microservices according to claim 6, characterized in that, The process of determining the microservice recommendation list corresponding to the target user based on the target microservice set and the candidate feature vector set includes: Based on the candidate feature vector set, a candidate microservice set is determined. Duplicate microservices in the target microservice set and the candidate microservice set are removed. The remaining microservices in the target microservice set and the candidate microservice set are determined as the microservice recommendation list corresponding to the target user.

8. An intelligent recommendation device for campus microservices, characterized in that, include: The acquisition module is used to acquire metadata from multiple microservices, vectorize the preprocessed metadata to obtain multiple feature vectors, and construct a microservice feature matrix based on the multiple feature vectors. The generation module is used to obtain the initial query statement of the target user, and generate multiple query vectors based on the initial query statement, the target user's historical behavior, and the service association rule base. The service association rule base is constructed based on the historical behavior of multiple users. The determination module is used to determine the candidate feature vector set in the microservice feature matrix based on multiple query vectors, and to determine the microservice recommendation list corresponding to the target user based on the target user's historical behavior, the historical behavior of multiple users, and the candidate feature vector set.

9. A recommended device, characterized in that, Including memory and processor, among which, The memory is used to store programs; The processor, coupled to the memory, is used to execute the program stored in the memory to implement the steps in the intelligent recommendation method for campus microservices as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, Used to store computer-readable programs or instructions, which, when executed by a processor, can implement the steps in the intelligent recommendation method for campus microservices as described in any one of claims 1 to 7.

Images