Internet information service method and system based on cloud computing

Abstract

The application discloses an internet information service method and system based on cloud computing, comprising: dynamically allocating cloud computing resources based on the number of user requests obtained in real time, and constructing a distributed processing cluster; performing protocol analysis and feature extraction on the requests, and obtaining a feature dataset; performing preliminary screening of all network information based on the request types in the feature dataset, and obtaining a preliminary screening result set containing at least three keyword matches; constructing an internet information matching model, performing feature matching degree calculation on the preliminary screening result set and the request data based on the preliminary screening result set, and performing rescreening based on the matching result; selecting the top N high matching degree results to form a query result based on the rescreening result, and generating a response data packet containing an abstract, source credibility and a preview interface; the application has the advantages that dynamic resource allocation, accurate request analysis, intelligent information screening and a deep matching model are used, system performance and response accuracy are optimized, and efficient and reliable services are provided.

Description

Technical Field

[0001] This invention relates to information service technology, and in particular to a cloud computing-based Internet information service method and system. Background Technology

[0002] Internet information services encompass multiple fields such as data processing, matching algorithms, and information filtering, aiming to provide users with accurate and personalized content recommendations through intelligent technologies. With the rapid development of the internet, the amount of information is growing exponentially, and traditional manual filtering and simple keyword searches can no longer meet the demand for efficient and accurate information.

[0003] Current internet information service methods lack flexibility in resource allocation, typically relying on static or pre-set resource configurations. They cannot dynamically adjust based on real-time user requests and system load, potentially leading to performance bottlenecks or resource waste under high concurrency. Many existing internet information service methods are simplistic in request parsing and feature extraction, often lacking in-depth analysis and failing to accurately extract core request features. This results in inefficient information filtering and matching, hindering rapid responses to specific user needs. Traditional filtering and matching mechanisms largely rely on basic keyword matching, lacking multi-level feature matching and intelligent re-filtering, leading to low relevance and accuracy, and failing to provide users with the most relevant information. Many existing methods lack systematic automatic scaling and load balancing mechanisms, making it difficult to handle large-scale, high-concurrency requests, easily resulting in slow system response and insufficient processing capacity. Summary of the Invention

[0004] To improve existing internet information service methods and systems, this paper proposes a cloud computing-based internet information service method and system. This method optimizes system performance by dynamically allocating resources and building distributed clusters, ensuring efficient processing of large-scale user requests. Through precise request parsing, feature matching, and re-filtering, it provides highly matched and reliable query results, thereby improving the accuracy and response speed of information services.

[0005] To achieve the above objectives, the technical solution adopted by the present invention is as follows:

[0006] Cloud computing-based internet information service methods include:

[0007] Dynamically allocate cloud computing resources based on the real-time number of user requests to build a distributed processing cluster;

[0008] Based on the Internet information service requests sent by the client, the requests are parsed and features are extracted to obtain a feature dataset;

[0009] Based on the request type in the feature dataset, perform preliminary screening of information across the entire network to obtain a preliminary screening result set containing at least three keyword matches;

[0010] Construct an Internet information matching model, calculate the feature matching degree between the initial screening result set and the request data, and further screen based on the matching results;

[0011] Based on the re-filtering results, the top N high-matching results are selected to form the query results, and a response data packet containing a summary, source credibility, and preview interface is generated;

[0012] The load balancer returns response packets to the client and records service logs.

[0013] Preferably, the step of dynamically allocating cloud computing resources based on the real-time acquired number of user requests to construct a distributed processing cluster specifically includes:

[0014] By deploying a monitoring system, we can collect real-time data on the number of user requests, response time, and system load. Based on log records, we can process and analyze the log data to obtain information on system load.

[0015] Design a distributed cluster architecture based on containerization technology, and decompose the request business into microservices to handle the distributed load on different nodes.

[0016] Set an auto-scaling strategy based on the number of requests or CPU and memory usage. When the number of requests increases, automatically start new virtual machines or container instances, and automatically reduce resources when the load decreases.

[0017] Adjust the load balancing algorithm based on different load conditions and select the most suitable allocation strategy, including round-robin, weighted round-robin, and least connections.

[0018] Preferably, the step of performing protocol parsing and feature extraction on the Internet information service request sent by the client to obtain the feature dataset specifically includes:

[0019] Based on the Internet information service request sent by the client, identify the protocol type of the request, and parse out the various contents of the request according to the protocol specification;

[0020] Feature extraction is performed on the request data after protocol parsing, including URL information, request header information, request method path, timestamp, request frequency, and request body content;

[0021] Based on the acquired feature data, it is transformed into structured data or feature vectors, and a request data feature set is constructed.

[0022] Preferably, the preliminary screening of information across the entire network based on the request type in the feature dataset, to obtain a preliminary screening result set containing at least three keyword matches, specifically includes:

[0023] Based on the request types in the feature dataset, a keyword set is defined, and a request type keyword library is constructed.

[0024] Based on the relevance of keywords to relevant fields in relation to request types, different weights are assigned to words.

[0025] Match relevant fields in the request feature data with domain keywords to remove requests that do not contain keywords or whose keyword matching is insufficient;

[0026] Request data involving at least three keywords is filtered out by inclusion matching, frequency matching, and context matching, and an initial screening result set is generated.

[0027] Preferably, the construction of the Internet information matching model, based on the initial screening result set, calculates the feature matching degree between it and the request data, and performs further screening based on the matching results, specifically includes:

[0028] Based on historical internet information service data, an internet information matching model is constructed by training a deep neural network model.

[0029] Input the feature vector data from the initial screening result set and the user request feature vector data into the Internet information matching model;

[0030] Obtain information data that matches the requested data information from the domain information of the initial screening result set, and perform multi-level screening based on the different domains to which the requested data information belongs to obtain multi-level matching information data;

[0031] Based on matching information data at all levels, weights are assigned according to the relevance of the domain, the overall matching degree is calculated and sorted, and the information data most relevant to the requested data is selected.

[0032] Preferably, the step of selecting the top N high-matching results based on the re-filtering results to form the query results, and generating a response data packet containing a summary, source credibility, and preview interface specifically includes:

[0033] Based on the re-filtering results, the top N high-matching results are selected to form the query results. The value of N is determined by the system response time and data dispersion, and the range of N is 3-10.

[0034] Based on the selected top N high-match results, extract key information from each result and generate a summary;

[0035] Based on the source identifiers of the information extracted from the matching results, calculate the source credibility score for each result;

[0036] Based on the selected N results, generate a preview interface that can quickly display the content;

[0037] The response data packet structure is constructed based on the query results, containing all the key information mentioned above, and the generated response data packet is returned to the client.

[0038] Furthermore, a cloud computing-based internet information service system is proposed, including:

[0039] The dynamic resource allocation module is mainly used to dynamically adjust cloud computing resources and build a distributed processing cluster based on real-time information such as the number of user requests and system load.

[0040] Request Protocol Parsing and Feature Extraction Module: The request protocol parsing and feature extraction module is mainly used to parse the requests sent by the client, extract feature data, and convert them into structured data;

[0041] Preliminary screening result set generation module: The preliminary screening result set generation module is mainly used to filter out a preliminary screening result set containing at least three keywords from the information of the entire network based on the feature type of the requested data, through keyword matching, frequency matching, context matching and other methods.

[0042] Information matching and re-filtering module: The information matching and re-filtering module is mainly used to construct an Internet information matching model, calculate the feature matching degree of the initial screening result set, and perform re-filtering based on the matching degree;

[0043] Query result generation module: The query result generation module is mainly used to generate a response data packet containing a summary, source credibility, and preview interface based on the re-filtering results;

[0044] Processor: The processor is mainly used for the calculation process of each formula and the construction calculation process of each model.

[0045] Compared with the prior art, the advantages of the present invention are:

[0046] By monitoring user request volume and system load in real time, the system automatically adjusts cloud computing resources to ensure efficient operation. Furthermore, by building a distributed cluster using containerization technology, it achieves load balancing and elastic resource scaling, preventing system overload or resource waste. In the request parsing and feature extraction process, this method accurately obtains the core data of the request by parsing the client request protocol and extracting features, transforming it into structured data and improving processing efficiency. For information filtering, based on the feature dataset, keyword matching and multi-level feature matching effectively filter out the most relevant information, ensuring the accuracy and relevance of the query results. By constructing an internet information matching model, the initial screening results are deeply analyzed and further filtered to improve matching accuracy, ensuring that the returned query results have high credibility, a summary and preview interface, and provide a fast and efficient information response service. Attached Figure Description

[0047] Figure 1 This is a schematic diagram of the method proposed in this invention;

[0048] Figure 2 This is a schematic diagram of cloud computing resource allocation proposed in this invention;

[0049] Figure 3 This is a schematic diagram illustrating the request parsing proposed in this invention;

[0050] Figure 4 This is a schematic diagram of the initial information screening proposed in this invention;

[0051] Figure 5 This is a schematic diagram illustrating the information re-filtering proposed in this invention;

[0052] Figure 6 This is a schematic diagram of the response data packet generation proposed in this invention;

[0053] Figure 7 This is an architecture diagram of the electronic devices in this solution;

[0054] Figure 8 This is a schematic diagram of the computer-readable storage medium structure in this scheme. Detailed Implementation

[0055] The following description is intended to disclose the invention and enable those skilled in the art to implement it. The preferred embodiments described below are merely examples, and other obvious variations will occur to those skilled in the art.

[0056] Cloud-based internet information service systems include:

[0057] The dynamic resource allocation module is mainly used to dynamically adjust cloud computing resources and build a distributed processing cluster based on real-time information such as the number of user requests and system load.

[0058] Request Protocol Parsing and Feature Extraction Module: The request protocol parsing and feature extraction module is mainly used to parse the requests sent by the client, extract feature data, and convert them into structured data;

[0059] Preliminary screening result set generation module: The preliminary screening result set generation module is mainly used to filter out a preliminary screening result set containing at least three keywords from the information of the entire network based on the feature type of the requested data, through keyword matching, frequency matching, context matching and other methods.

[0060] Information matching and re-filtering module: The information matching and re-filtering module is mainly used to construct an Internet information matching model, calculate the feature matching degree of the initial screening result set, and perform re-filtering based on the matching degree;

[0061] Query result generation module: The query result generation module is mainly used to generate a response data packet containing a summary, source credibility, and preview interface based on the re-filtering results;

[0062] Processor: The processor is mainly used for the calculation process of each formula and the construction calculation process of each model.

[0063] See Figure 1 As shown, the cloud computing-based internet information service method includes:

[0064] Step 1: Dynamically allocate cloud computing resources based on the real-time number of user requests to build a distributed processing cluster;

[0065] Step 2: Based on the Internet information service request sent by the client, perform protocol parsing and feature extraction on the request to obtain a feature dataset;

[0066] Step 3: Based on the request type in the feature dataset, perform preliminary screening of information across the entire network to obtain a preliminary screening result set containing at least three keyword matches;

[0067] Step 4: Construct an internet information matching model. Based on the initial screening result set, calculate the feature matching degree between it and the request data, and then perform further screening based on the matching results.

[0068] Step 5: Based on the re-filtering results, select the top N high-matching results to form the query results, and generate a response data packet containing a summary, source credibility, and preview interface;

[0069] Step 6: Return the response data packet to the client through the load balancer and record the service log.

[0070] See Figure 2 As shown, dynamically allocating cloud computing resources based on the real-time number of user requests to build a distributed processing cluster specifically includes:

[0071] By deploying a monitoring system, we can collect real-time data on the number of user requests, response time, and system load. Based on log records, we can process and analyze the log data to obtain information on system load.

[0072] Design a distributed cluster architecture based on containerization technology, and decompose the request business into microservices to handle the distributed load on different nodes.

[0073] Set an auto-scaling strategy based on the number of requests or CPU and memory usage. When the number of requests increases, automatically start new virtual machines or container instances, and automatically reduce resources when the load decreases.

[0074] Adjust the load balancing algorithm based on different load conditions and select the most suitable allocation strategy, including round-robin, weighted round-robin, and least connections.

[0075] Specifically, deploying a monitoring system is crucial to ensuring the healthy operation of the system. The data that the system needs to collect includes:

[0076] User request count: requests per second; response time: the latency between request and response, usually tracked through application layer logs; system load data: CPU utilization, memory utilization, disk I / O, network bandwidth, etc.

[0077] Containerization technology is used to design a scalable distributed cluster architecture, which breaks down the application into microservices and uses Docker to containerize each microservice. The containers are orchestrated and managed. Based on business logic and performance requirements, requests are split into multiple microservices, each of which is deployed as an independent container and runs on different nodes to achieve service discovery and load balancing.

[0078] Load balancing algorithms determine how to select the most suitable allocation strategy based on different load conditions. Common load balancing algorithms include round-robin, weighted round-robin, and least connections.

[0079] In the round-robin algorithm, requests are distributed to different servers in turn in a sequential manner, which is suitable for scenarios where each server has similar processing capabilities;

[0080] In the weighted round-robin algorithm, each server allocates requests according to its weight, and servers with higher weights receive more requests.

[0081] In the least connections algorithm, requests are assigned to the server with the fewest current connections, making it suitable for handling situations where requests take a long time or the load is uneven.

[0082] See Figure 3 As shown, based on the Internet information service request sent by the client, the request is parsed and features are extracted to obtain the feature dataset. Specifically, this includes:

[0083] Based on the Internet information service request sent by the client, identify the protocol type of the request, and parse out the various contents of the request according to the protocol specification;

[0084] Feature extraction is performed on the request data after protocol parsing, including URL information, request header information, request method path, timestamp, request frequency, and request body content;

[0085] Based on the acquired feature data, it is transformed into structured data or feature vectors, and a request data feature set is constructed.

[0086] Specifically, client requests usually include a protocol field, and the protocol type is determined based on the information in the protocol header, such as HTTP, HTTPS, gRPC, etc.

[0087] After the protocol is parsed, the request content needs to be parsed out, including URL information, request header information, request method path, timestamp, and request body content. The URL contains the target resource address of the request. The URL in an HTTP request usually contains the following parts: protocol, host, port number, path, and query parameters.

[0088] Based on the request data after protocol parsing, key features are extracted for analysis and transformed into structured data or feature vectors. Through feature vectorization methods, the extracted features are converted into data structures that can be used by machine learning models, for example:

[0089] Numerical features: directly converted into numerical values, such as path length, request frequency, and request body size;

[0090] Categorical features: One-hot encoding is used to convert protocol type, request method, etc., into numerical vectors;

[0091] Text features: If the request body contains text data, convert it into numerical features using TF-IDF or Word2Vec methods.

[0092] See Figure 4 As shown, based on the request type in the feature dataset, a preliminary screening of information across the entire network is performed to obtain a preliminary screening result set containing at least three keyword matches. Specifically, this includes:

[0093] Based on the request types in the feature dataset, a keyword set is defined, and a request type keyword library is constructed.

[0094] Based on the relevance of keywords to relevant fields in relation to request types, different weights are assigned to words.

[0095] Match relevant fields in the request feature data with domain keywords to remove requests that do not contain keywords or whose keyword matching is insufficient;

[0096] Request data involving at least three keywords is filtered out by inclusion matching, frequency matching, and context matching, and an initial screening result set is generated.

[0097] Specifically, different request types are usually associated with specific keywords. Each request type and its corresponding keyword set are stored in a mapping structure, such as a dictionary or database. Each request type's keyword set may have different weight values, which represent the relevance of the keyword to the request type. The higher the weight, the more critical the keyword.

[0098] Each keyword is assigned a different weight based on its relevance to a specific request type. The weighting is based on the following factors:

[0099] Domain relevance: The frequency and importance of certain keywords in a specific domain determine their weight;

[0100] Keyword frequency: The higher the frequency of a keyword in historical data, the more important it may be.

[0101] The formula is:

[0102] w ij =f ij ×r ij

[0103] Among them, w ij To obtain the weight of keyword j in request i, f ij r represents the frequency of keyword i in request j. ij The relevance weight of keyword i in request type j;

[0104] The relevant fields in the request feature data are matched with keywords. First, the HTTP method of the request is matched with the keyword set defined in the keyword library; the path part is extracted from the URL of the request and matched according to the keywords in the path; if the request method is POST or PUT, the keywords are extracted from the request body and matched.

[0105] Specifically, the matching function checks whether a field in the request contains a keyword from the keyword library.

[0106] Frequency matching specifically determines the importance of keywords by calculating the frequency of their appearance in a request;

[0107] Context matching specifically involves checking whether the context of the request supports the appearance of keywords, such as by checking the combination of the request method and the request body content.

[0108] By matching keywords and filtering request data containing at least three keywords based on matching frequency and context, an initial screening result set is generated.

[0109] See Figure 5 As shown, an internet information matching model is constructed. Based on the initial screening result set, the feature matching degree is calculated between the model and the request data. Further screening is then performed based on the matching results. Specifically, this includes:

[0110] Based on historical internet information service data, an internet information matching model is constructed by training a deep neural network model.

[0111] Input the feature vector data from the initial screening result set and the user request feature vector data into the Internet information matching model;

[0112] Obtain information data that matches the requested data information from the domain information of the initial screening result set, and perform multi-level screening based on the different domains to which the requested data information belongs to obtain multi-level matching information data;

[0113] Based on matching information data at all levels, weights are assigned according to the relevance of the domain, the overall matching degree is calculated and sorted, and the information data most relevant to the requested data is selected.

[0114] Specifically, historical data is transformed into a format acceptable to the model to generate a training dataset. Each data point contains features of user requests and corresponding service data. The input layer accepts all feature vectors, including various features of requests and services. The hidden layer performs feature extraction and non-linear mapping through multiple fully connected layers. The output layer generates predicted values.

[0115] The feature vectors of each request data and the feature vectors of the service data may have different dimensions and types. The category features are processed by embedding layers or one-hot encoding so that all features are on the same scale. The feature vectors of the request data and the feature vectors of the service data are concatenated and used as the input of the model.

[0116] Based on the domain information in the initial screening result set, service data that matches the request data is obtained. For example, if the request type is an API request, only information related to the API service will be considered for matching;

[0117] Through a multi-layered filtering mechanism, the most relevant information data is further refined and selected. This process is based on the domain of the requested data and its similarity to the service data.

[0118] First-level filtering: Filtering service data related to the request based on the basic information of the request.

[0119] The second layer of filtering: further filtering of service data that meet the criteria based on request body content, timestamp, request frequency, etc.

[0120] The third layer of filtering: Select the best service based on the response time, status, and other factors of the service data;

[0121] By analyzing the service data selected at each level, different weights are assigned based on the relevance of the domain, and the overall matching degree is calculated using the following formula:

[0122]

[0123] Among them, matching_score i For the matching degree of the i-th service data, w domain Domain weights;

[0124] All service data are sorted based on overall matching degree, and the service data most relevant to the requested data is selected.

[0125] See Figure 6 As shown, based on the re-filtering results, the top N high-matching results are selected to form the query results, and a response data packet containing a summary, source credibility, and preview interface is generated. Specifically, this includes:

[0126] Based on the re-filtering results, the top N high-matching results are selected to form the query results. The value of N is determined by the system response time and data dispersion, and the range of N is 3-10.

[0127] Based on the selected top N high-match results, extract key information from each result and generate a summary;

[0128] Based on the source identifiers of the information extracted from the matching results, calculate the source credibility score for each result;

[0129] Based on the selected N results, generate a preview interface that can quickly display the content;

[0130] The response data packet structure is constructed based on the query results, containing all the key information mentioned above, and the generated response data packet is returned to the client.

[0131] Specifically, the selection of N is based on the following two factors:

[0132] System response time: A lower response time allows more results to be returned, while a higher response time results in fewer results.

[0133] Data dispersion: When the data is highly discrete, it may be necessary to limit the number of N to avoid returning too many irrelevant results;

[0134] The formula is:

[0135]

[0136] Among them, T threshold T is the maximum threshold for system response time. response The response time for the current query is given by `max_results`, which is the maximum value of N, usually 10. `min` and `round` are limits to ensure that N is within a reasonable range.

[0137] For each selected query result, we need to extract key information, which may include:

[0138] Title: A brief text describing the query; Summary: A concise summary of the query results; Timestamp: The time the service responded or the time the data was created; Related fields: For example, related URLs, IDs, tags, etc.

[0139] There are generally two ways to generate summaries:

[0140] Rule-based summary generation: Extract the first few sentences, key fields, etc. from the query results and then concatenate them;

[0141] Model-based summary generation: using natural language processing models, such as BERT, to generate more complex and coherent summaries.

[0142] Each query result is assigned a credibility score based on the credibility of its source. At the same time, a concise preview interface is generated based on the selected N results. This interface includes a summary, key information, and credibility score for each result.

[0143] Furthermore, the method according to the embodiments of this application can also be achieved by means of... Figure 7 The architecture of the electronic device shown is used to implement this. For example... Figure 7 As shown, the electronic device 500 may include a bus 501, one or more CPUs 502, a read-only memory (ROM) 503, a random access memory (RAM) 504, a communication port 505 connected to a network, an input / output component 506, a hard disk 507, etc. The storage device in the electronic device 500, such as the ROM 503 or the hard disk 507, may store the cloud computing-based Internet information service method and system provided in this application. The electronic device 500 may also include a terminal interface 508. Of course, Figure 7 The architecture shown is merely exemplary and can be omitted as needed when implementing different devices. Figure 7 One or more components in the illustrated electronic device.

[0144] Figure 8 This is a schematic diagram of a computer-readable storage medium structure provided in one embodiment of this application. Figure 8The diagram illustrates a computer-readable storage medium 600 according to one embodiment of this application. The computer-readable storage medium 600 stores computer-readable instructions. When executed by a processor, the computer-readable instructions can perform the cloud computing-based internet information service method and system according to an embodiment of this application, as described with reference to the above figures. The storage medium 600 includes, but is not limited to, volatile memory and / or non-volatile memory. Volatile memory may include, for example, random access memory (RAM) and cache memory. Non-volatile memory may include, for example, read-only memory (ROM), hard disk, flash memory, etc.

[0145] It should be noted that the order of the above embodiments of the present invention is merely for descriptive purposes and does not represent the superiority or inferiority of the embodiments. Furthermore, the above description focuses on specific embodiments of this specification. Additionally, the processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired results. In some embodiments, multitasking and parallel processing are possible or may be advantageous.

[0146] The various embodiments in this specification are described in a progressive manner. The same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on describing the differences from other embodiments.

[0147] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A cloud computing-based Internet information service method, characterized by, The application comprises the following steps: Dynamically allocate cloud computing resources based on the number of user requests obtained in real time to build a distributed processing cluster; Based on the internet information service request sent by the client, protocol analysis and feature extraction are performed on the request to obtain a feature dataset; Based on the request type in the feature dataset, preliminary screening of the network information is performed to obtain a preliminary screening result set containing at least three keyword matches; Based on historical internet information service data, an internet information matching model is constructed through deep neural network model training; The feature vector data in the preliminary screening result set and the user request feature vector data are input into the internet information matching model; Information data matching the request data information in the domain information of the preliminary screening result set is obtained, and multi-level screening is performed based on different domains to which the request data information belongs to obtain multi-level matching information data; Based on all levels of matching information data, weights are assigned according to the relevance of the domains, the comprehensive matching degree is calculated and sorted, and the information data most relevant to the request data is screened out; Based on the re-screening result, the top N high matching degree results are selected to form the query result, and a response data packet containing an abstract, source credibility, and a preview interface is generated; The response data packet is returned to the client through the load balancer, and service logs are recorded.

2. The cloud-computing-based Internet information service method according to claim 1, characterized in that, The dynamic allocation of cloud computing resources based on the number of user requests obtained in real time to build a distributed processing cluster specifically comprises the following steps: Collect real-time user request quantity, response time, and system load data through a deployment monitoring system, process and analyze the log data based on log records, and obtain the system load condition; Design a distributed cluster architecture based on containerization technology, and split the request business into microservices to handle scattered loads on different nodes; Set an automatic expansion strategy based on the number of requests or CPU and memory usage, automatically start new virtual machines or container instances when the request volume increases, and automatically reduce resources when the load decreases; Adjust the load balancing algorithm based on different load conditions to select the most suitable distribution strategy, including round robin, weighted round robin, and least connection. 3.The cloud-computing-based Internet information service method according to claim 1, characterized in that, The protocol analysis and feature extraction of the request based on the internet information service request sent by the client to obtain a feature dataset specifically comprises the following steps: Based on the internet information service request sent by the client, identify the protocol type of the request and parse the contents of the request according to the protocol specification; Based on the request data after protocol analysis, perform feature extraction, including URL information, request header information, request method path, timestamp, request frequency, and request body content; Based on the obtained feature data, convert it into structured data or feature vectors, and construct a request data feature set.

4. The cloud-computing-based Internet information service method according to claim 1, characterized in that, The preliminary screening of the network information based on the request type in the feature dataset to obtain a preliminary screening result set containing at least three keyword matches specifically comprises the following steps: Based on the request type in the feature dataset, define a keyword set and construct a request type keyword library; Based on the relevance of the keywords and the related domains of the request type, assign different weights to the keywords; Match the relevant fields in the request feature data with the domain keywords, and remove requests that do not contain keywords or have insufficient keyword matching. The request data involving at least three keywords is screened out through matching, frequency matching and context matching, and a preliminary screening result set is generated. 5.The cloud-computing-based Internet information service method according to claim 1, characterized in that, Based on the re-screening result, the top N high matching degree results are selected to form a query result, and a response data packet including an abstract, a source credibility and a preview interface is generated. Based on the re-screening result, the top N high matching degree results are selected to form a query result, and a response data packet including an abstract, a source credibility and a preview interface is generated. Based on the re-screening result, the top N high matching degree results are selected to form a query result, and a response data packet including an abstract, a source credibility and a preview interface is generated. Based on the re-screening result, the top N high matching degree results are selected to form a query result, and a response data packet including an abstract, a source credibility and a preview interface is generated. Based on the re-screening result, the top N high matching degree results are selected to form a query result, and a response data packet including an abstract, a source credibility and a preview interface is generated. Based on the re-screening result, the top N high matching degree results are selected to form a query result, and a response data packet including an abstract, a source credibility and a preview interface is generated.

6. A cloud computing-based Internet information service system for implementing the cloud computing-based Internet information service method according to any one of claims 1 to 5, characterized by Based on the query result, a response data packet structure is constructed, including all the above key information, and the generated response data packet is returned to the client. It comprises: The resource dynamic allocation module is mainly used for dynamically adjusting cloud computing resources according to real-time acquired user request quantity, system load and other information, and constructing a distributed processing cluster. The request protocol analysis and feature extraction module is mainly used for analyzing the request sent by the client, extracting feature data, and converting it into structured data. The preliminary screening result set generation module is mainly used for screening a preliminary screening result set containing at least three keywords from the information on the Internet based on the feature type of the request data through keyword matching, frequency matching and context matching. The information matching and re-screening module is mainly used for calculating the feature matching degree of the preliminary screening result set by constructing an Internet information matching model, and re-screening according to the matching degree. The query result generation module is mainly used for generating a response data packet including an abstract, a source credibility and a preview interface based on the re-screening result.

7. An electronic device, comprising: The processor is mainly used for the calculation process of each formula and the construction and calculation process of each model. It comprises: At least one processor; And a memory in communication connection with the at least one processor; wherein 8. A computer-readable storage medium storing computer-readable instructions, the computer-readable instructions comprising: The memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to execute the cloud computing-based Internet information service method of any one of claims 1-5. The computer readable instructions are executed by the processor to implement the cloud computing-based Internet information service method of any one of claims 1-5.

Images