Smart campus community data storage method and system

By calculating the importance of information on the smart campus community platform and managing different storage durations, the problem of data capacity is solved, efficient storage and management of information is achieved, and campus management efficiency is improved.

CN118964516AInactive Publication Date: 2024-11-15JINAN GAOPIN WEIYE INFORMATION TECH CO LTD
View PDF 9 Cites 0 Cited by

Patent Information

Application Number
CN202411447906.8
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-10-17
Publication Date
2024-11-15
Estimated Expiration
Not applicable · inactive patent

AI Technical Summary

Technical Problem

The campus community platform data capacity in the smart campus system is full, making it difficult to store new data and cannot meet data management needs.

Method used

By obtaining the target information of the smart campus community platform, calculating its importance, and managing the information with different storage durations according to the degree of importance, including permanent storage and temporary storage, avoiding insufficient storage capacity and unnecessary data cleaning.

Benefits of technology

Effectively manage information storage, avoid insufficient storage capacity and data loss, improve information value mining and campus management efficiency, and provide strong support for smart campus management.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118964516A_ABST
    Figure CN118964516A_ABST
Patent Text Reader

Abstract

The invention relates to the field of electric data processing, in particular to a smart campus community data storage method and system, and the method comprises the steps: obtaining release data of target information released in a smart campus community platform, the release data comprising release content and release time of the target information; calculating the importance of the target information according to the published data; the target information is stored according to the importance degree of the target information, and different importance degrees correspond to different storage durations. And performing importance analysis on the information according to the release data of the information, and storing the information with different importance degrees in different storage durations according to the importance degrees of the information. Therefore, the problem of insufficient storage capacity caused by permanent storage of all the information can be avoided, and inconvenience of deleting important data caused by regular cleaning of all the information can also be avoided.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the field of data processing, and in particular to a smart campus community data storage method and system. Background Art

[0002] The smart campus system is an important part of the current education informatization and smart campus construction. Its technical background is rooted in the rapid development of modern information technologies such as big data, cloud computing, the Internet of Things and artificial intelligence. With the continuous improvement of the degree of informatization, the amount of data generated by the smart campus system has increased dramatically. These data include but are not limited to student learning data, campus security monitoring data, campus resource usage data, teacher-student interaction data, etc. A large amount of data is transmitted through the campus network and transferred to the smart campus system for storage so that data can be consulted when needed.

[0003] With the popularization of smart campus systems, the campus community platform in the smart campus system has become an important platform for students to communicate, learn and interact. A large amount of information (such as event notifications, employment information, etc.) from various subjects (such as individual students, student departments, etc.) is generated on the smart campus community platform every day. These various information are transmitted to the smart campus community platform by different subjects through different terminals and different data channels, which can easily lead to the data capacity of the smart campus community platform being full, making it difficult to continue to store new data and unable to meet the data management needs of the smart campus system. Summary of the invention

[0004] In order to solve the above technical problems, the object of the present invention is to provide a smart campus community data storage method, comprising: Acquire the publishing data of the target information published in the smart campus community platform, wherein the publishing data includes the publishing content and publishing time of the target information; Calculating the importance of the target information according to the published data; The target information is stored according to the importance of the target information, wherein different importance levels correspond to different storage durations.

[0005] Optionally, the published data further includes a publishing subject, and the calculating the importance of the target information according to the published data includes: Determining the target category to which the target information belongs according to the publisher of the target information; The importance of the target information is calculated according to the publishing content and publishing time of the target information and the publishing data of other information in the target category.

[0006] Optionally, calculating the importance of the target information according to the release content and release time of the target information and release data of other information in the target category includes: Calculate the similarity between the target information and other information in the target category based on the published content of all information in the target category; Calculate the standardization of the target information according to the similarity between the target information and other information in the target category, the proportion of preset words in the published content of the target information, and the standardization of the content of the target information; Calculate the periodicity of the target information according to the standardization of the target information and the release time of the target information; The importance of the target information is calculated according to the periodicity of the target information and the importance of the target category.

[0007] Optionally, the calculating the similarity between the target information and other information in the target category according to the published contents of all information in the target category includes: Numerically processing the published content of all information in the target category to obtain multiple coding curves, wherein one piece of information corresponds to one coding curve, and the target information corresponds to the target coding curve; The similarity between the target information and other information in the target category is calculated according to the length and browsing time of the target coding curve and the length and browsing time of the coding curve of other information in the target category.

[0008] Optionally, the publishing contents of all information in the target category are numerically processed to obtain multiple coding curves, including: For each piece of information, use text embedding technology to convert the text in the information release content into multiple word vectors; The multiple word vectors are positionally encoded according to the positions of different sentences in the information to obtain corresponding encoding curves.

[0009] Optionally, before calculating the standardization of the target information, the method further includes: Calculating the content standardization of the target information according to the first Euclidean distance and the second Euclidean distance; The first Euclidean distance is, in the first cluster, an average value of the Euclidean distances between the word vector at the first position in the target information and the word vector at the corresponding position in other information in the target category; The second Euclidean distance is the average value of the Euclidean distances between the word vector at the second position in the target information and the word vector at the corresponding position in other information in the target category in the second cluster; The word vector at the first position includes the word vector located in the front of a first preset number of positions and the word vector located in the back of a second preset number of positions, and the word vector at the second position includes the word vectors in the word vector sequence except the word vector at the first position.

[0010] Optionally, before calculating the content standardization of the target information according to the first Euclidean distance and the second Euclidean distance, the method includes: Convert all information in the target category into multiple word vector sequences, where one piece of information corresponds to one word vector sequence; Match each word vector sequence in the target category with the standard word vector sequence to obtain multiple matching word vectors; The multiple word vector sequences are clustered according to the normalized distance of all matching word vectors to obtain a first cluster and a second cluster, wherein the first cluster includes word vectors that are in the first position in each word vector sequence and whose normalized distance of matching word vectors is less than a preset distance; and the second cluster includes word vectors in the multiple word vector sequences except the word vectors included in the first cluster.

[0011] Optionally, the calculating the periodicity of the target information according to the standardization of the target information and the release time of the target information includes: Sorting all the information in the target category according to the release time to obtain a release time sequence of all the information in the target category; Construct a plane with the standardization as the vertical axis and the release time series as the horizontal axis, map all the information in the target category to the plane respectively, and divide it into periods with a preset time length to obtain multiple fluctuation curves, wherein all the information in the same period is connected into one fluctuation curve; The periodicity of the target information is calculated based on the standardization of the target information, the standardization of the standard information, and the similarity between the fluctuation curve of the target information and the fluctuation curve of the standard information, wherein the standard information is information whose release time in other periods matches the release time of the target information.

[0012] Optionally, calculating the importance of the target information according to the periodicity of the target information and the importance of the target category includes: Calculate the average value of the periodicity of all information in the target category as the importance of the target category; The product of the importance of the target category and the periodicity of the target information is calculated as the importance of the target information.

[0013] The present invention also provides a smart campus community data storage system, the system comprising: An acquisition module, used to acquire the publishing data of the target information published in the smart campus community platform, wherein the publishing data includes the publishing content and publishing time of the target information; A calculation module, used for calculating the importance of the target information according to the published data; The storage module is used to store the target information according to the importance of the target information, wherein different importance levels correspond to different storage durations.

[0014] The present invention has the following beneficial effects: the importance of information is analyzed according to the information release data, and information of different importance is stored in different storage methods according to the importance of the information, for example, information of higher importance can be permanently stored, and information of lower importance can be temporarily stored. This can avoid the problem of insufficient storage capacity caused by permanently storing all information, and the inconvenience of deleting important data caused by regularly cleaning up all information, and can also more effectively mine and utilize the value of information, thereby improving campus management efficiency and providing strong support for smart campus management. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] In order to more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings required for use in the embodiments or the prior art descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0016] Figure 1 One of the flow charts of the smart campus community data storage method provided by an embodiment of the present invention; Figure 2 A second flow chart of the smart campus community data storage method provided by an embodiment of the present invention; Figure 3 A schematic diagram of the modules of the smart campus community data storage system provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0017] In order to further explain the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the following is a detailed description of the specific implementation method, structure, features and effects of a smart campus community data storage method and system proposed by the present invention in combination with the accompanying drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" does not necessarily refer to the same embodiment. In addition, specific features, structures or characteristics in one or more embodiments may be combined in any suitable form.

[0018] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.

[0019] The present invention provides a smart campus community data storage method, and the scheme of the present invention is described in detail below with reference to the accompanying drawings.

[0020] See also Figure 1 , which shows one of the flow diagrams of a smart campus community data storage method provided by an embodiment of the present invention, the smart campus community data storage method (hereinafter referred to as "the method") comprises the following steps: Step 101: Acquire the publishing data of the target information published in the smart campus community platform, wherein the publishing data includes the publishing content and publishing time of the target information.

[0021] With the widespread use of the smart campus community platform, many daily activities of campus personnel rely on the platform. Various information will be posted on the smart campus community platform, such as notices, shared information, topic discussions, mutual assistance, community information, recruitment information, idle sales, campus consultation, etc. In this step, the method obtains the release data of the target information released in the smart campus community platform, and the release data includes the release content and release time of the target information. In some embodiments, the target information is text information. Specifically, the method can use web crawler technology to obtain the release data of information on the smart campus community platform.

[0022] Step 102: Calculate the importance of the target information according to the published data.

[0023] Generally speaking, the content and time of information release will affect the importance of the information. For example, the more standardized the content and the more periodic the release time, the higher the importance of the information. In this step, the method calculates the importance of the target information based on the release data of the target information. Specifically, the method can first calculate the similarity between the target information and other information based on the release content of all information on the smart campus community platform; then calculate the standardization of the target information based on the similarity between the target information and other information and the release content of the target information itself (such as the use of some special words in the release content, the standardization of the content, etc.); after obtaining the standardization of the target information, the method can further calculate the periodicity of the target information based on the standardization of the target information and the release time of the target information, and finally calculate the importance of the target information based on the obtained periodicity of the target information and the importance of other information on the smart campus community platform.

[0024] Step 103: Store the target information according to the importance of the target information, wherein different importance levels correspond to different storage durations.

[0025] After determining the importance of the target information, the method may store the target information according to the importance of the target information. Specifically, for information of different importance, the method may store the information in different storage modes, such as storing for different durations.

[0026] In this embodiment, the method can pre-acquire a mapping relationship table between different importance levels and different storage methods, wherein different importance levels correspond to different storage methods, and different storage methods have different storage durations. After calculating the importance of the target information, the method can determine its importance level (i.e., the importance range it falls into) according to the importance of the target information, and then determine the corresponding storage method from the mapping relationship table as the target storage method according to the importance of the target information, and finally use the target storage method to store the target information.

[0027] For example, Table 1 below is a mapping table of importance and storage method in one embodiment of the present invention. For information with a high degree of importance, it is permanently stored; for information with a medium degree of importance, it is automatically deleted after one year of storage; for information with a low degree of importance, it is automatically deleted after six months of storage. It should be noted that the mapping relationship in Table 1 below is only for illustration, but is not limited to this. It can be understood that in actual application, the degree of importance can be divided into more or fewer levels according to needs, and storage methods of different lengths can be provided for each level, or for information with a high degree of importance, it can be backed up while being permanently stored to avoid the loss of important information due to accidental deletion.

[0028] Table 1 Importance and storage method mapping table In this embodiment, the method obtains the publishing data of the target information published in the smart campus community platform, wherein the publishing data includes the publishing content and publishing time of the target information; calculates the importance of the target information according to the publishing data; and stores the target information according to the importance of the target information, wherein different importance levels correspond to different storage durations. The information is analyzed for importance according to the publishing data of the information, and information of different importance levels is stored in different storage methods according to the importance of the information. For example, information of higher importance can be permanently stored, and information of lower importance can be temporarily stored. This can avoid the problem of insufficient storage capacity caused by permanently storing all information, avoid the inconvenience of deleting important data caused by regularly cleaning up all information, and more effectively mine and utilize the value of information, thereby improving campus management efficiency and providing strong support for smart campus management.

[0029] See also Figure 2 , which shows a second flow chart of a smart campus community data storage method provided by an embodiment of the present invention, the method comprising the following steps: Step 201: Acquire the publishing data of the target information published in the smart campus community platform, wherein the publishing data includes the publishing subject, publishing content and publishing time of the target information.

[0030] In this step, the method obtains the publishing data of the target information published in the smart campus community platform. In addition to the publishing content and publishing time of the target information, the publishing data also includes the publishing subject of the target information, that is, the author who published the target information.

[0031] Step 202: Determine the target category to which the target information belongs according to the publisher of the target information.

[0032] It is understandable that for the information published on the smart campus community platform, when the publishing subject is different, the importance of the message may also be different. For example, when the publishing subject is official, its authority is higher, and accordingly, the importance of the message it publishes should also be higher. In this embodiment, the method can cluster all the information on the smart campus community platform according to the publishing subject to obtain several message categories and clusters corresponding to each message category, wherein different message categories correspond to different publishing subjects, and the cluster corresponding to each message category includes all messages published by the corresponding publishing subject, that is, one cluster includes all messages published by one publishing subject, and different clusters include messages published by different publishing subjects. The method can use web crawler technology to obtain the publishing subject of each message on the smart campus community platform, and then cluster them according to the publishing subject.

[0033] In this step, the method determines the target category to which the target information belongs according to the publisher of the target information, so as to subsequently further calculate the importance of the target message according to the target category to which it belongs.

[0034] Step 203: Calculate the importance of the target information according to the release content and release time of the target information and the release data of other information in the target category.

[0035] In addition to the publishing content and publishing time affecting the importance of information, other content published by the same publishing subject can also be used as one of the factors affecting the importance of information. In this step, the method calculates the importance of the target information based on the publishing content, publishing time and publishing data of other information in the target category to which the target information belongs.

[0036] Specifically, the method can calculate the similarity between the target information and other information in the target category based on the published content of all information in the target category; then calculate the standardization of the target information based on the similarity between the target information and other information in the target category and the published content of the target information itself (such as the use of some special words in the published content, the standardization of the content, etc.); then calculate the periodicity of the target information based on the standardization of the target information and the release time of the target information; finally, calculate the importance of the target information based on the periodicity of the target information and the importance of the target category. In this step, the method not only calculates the importance of the target information based on its published data, but also further combines the published data of other information in the target category to which it belongs to evaluate the importance of the target information in multiple dimensions.

[0037] Step 204: Store the target information according to the importance of the target information, wherein different importance levels correspond to different storage durations.

[0038] This step 204 is the same as step 103 in the previous embodiment and will not be described again here.

[0039] Optionally, the importance of the target information is calculated based on the published content, published time of the target information and published data of other information in the target category, including: calculating the similarity between the target information and other information in the target category based on the published content of all information in the target category; calculating the standardization of the target information based on the similarity between the target information and other information in the target category, the proportion of preset words in the published content of the target information and the content standardization of the target information; calculating the periodicity of the target information based on the standardization of the target information and the published time of the target information; calculating the importance of the target information based on the periodicity of the target information and the importance of the target category.

[0040] Taking into account that the degree of similarity between the target information and other information in the same category will also affect its importance, in this embodiment, the method can digitize all the information in the target category to obtain a coding curve for each information, and then calculate the similarity between the target information and other information in the same category based on the coding curve of the information and the browsing time, and then further calculate the importance of the target information based on the obtained similarity.

[0041] In addition to the similarity with other information in the same category, the information's own published content will also affect its importance. For example, the more standardized the information's published content is, the higher its importance may be. Accordingly, after obtaining the similarity between the target information and other information in the target category, the method also calculates the standardization of the target information based on the similarity and the situation of the target message's own published content. The target message's own published content may include the use of some special words (such as adjectives, adverbs, etc.) in the published content, the standardization of the content, etc.

[0042] For some periodic notices and announcements, they may be official notices with high importance. Therefore, the importance of the information can be evaluated in combination with the time regularity of the corresponding publishing subject releasing the information. Specifically, the method can calculate the periodicity of the target information according to the standardization of the target information and the release time of the target information, and after obtaining the periodicity of the target information, comprehensively calculate the importance of the target information in combination with the overall importance of the target category to which the target message belongs.

[0043] Optionally, the calculating the similarity between the target information and other information in the target category based on the published content of all information in the target category includes: digitizing the published content of all information in the target category to obtain multiple coding curves, wherein one piece of information corresponds to one coding curve, and the target information corresponds to a target coding curve; calculating the similarity between the target information and other information in the target category based on the length and browsing time of the target coding curve, as well as the length and browsing time of the coding curves of other information in the target category.

[0044] In this embodiment, the method can perform numerical processing on the published content of all information published on the smart campus community platform to obtain a coding curve corresponding to each information. Then, according to the length and browsing time of the target coding curve corresponding to the target information, and the length and browsing time of the coding curve of other information in the target category to which the target information belongs, the similarity between the target information and other information in the target category is calculated.

[0045] The digitization of the published content of the information may be performed by using a natural language processing (NLP) algorithm to digitize the text in the published content, such as using contextual embeddings to convert the text in the published content of the information into multiple word vectors, and then position encoding the multiple word vectors according to the positions of different sentences in the information to obtain a coding curve corresponding to the information.

[0046] The length of the coding curve will be affected by the published content, for example, the content format and text length of the published content will affect the length of the coding curve. Considering that the published information is not only affected by its own published data, but also more or less affected by the browsing status, therefore, in this embodiment, the method can combine the length of the coding curve and the browsing time of the information to calculate the similarity between the target information and other information in the same category.

[0047] Specifically, the following formula can be used to calculate the similarity between the target information and other information in the target category to which it belongs: : in, Indicates the target information Similarity to other information in the target category to which it belongs; Indicates the first The number of the message, Indicates the total amount of information in the target category; Indicates the target information The length of the corresponding target encoding curve, Indicates the target information Average browsing time of Indicates the first The length of the coding curve corresponding to the information, Indicates the first Average browsing time of each piece of information; Represents the normalization function.

[0048] Considering that the text length of some common information may be close to that of official information, in order to further distinguish the two, similarity calculation can be combined with browsing time. It is understandable that for official information, the browsing time is generally longer, so the browsing time is used as the weight to calculate the similarity between the information and other information in the same category.

[0049] In the above formula, Target information The comprehensive characteristics of The larger the target information The longer the average browsing time is), The larger the target information The higher the importance; For the target category The comprehensive characteristics of the information are immediately The larger the target category, the The longer the average browsing time of each piece of information is, The larger the target category, the The more important the information is.

[0050] Indicates target information The target category The difference in comprehensive feature performance between the pieces of information, The smaller the value, the more targeted the information. The target category The smaller the difference in comprehensive characteristics between pieces of information; Indicates target information The overall difference in the combined feature representation of all information in the target category, The smaller the value, the more targeted the information. The smaller the overall difference in the comprehensive feature performance between all information in the target category, the Similarity to other information in the target category The bigger it is.

[0051] Optionally, the published content of all information in the target category is numerically processed to obtain multiple encoding curves, including: for each information, using text embedding technology to convert the text in the published content of the information into multiple word vectors; position encoding the multiple word vectors according to the positions of different sentences in the information to obtain corresponding encoding curves.

[0052] In this embodiment, for each information in the target category, the method uses text embedding technology to convert the text in the published content of the information into multiple word vectors; the multiple word vectors are positionally encoded according to the positions of different sentences in the information to obtain the encoding curve corresponding to the information. The BERT (Bidirectional Encoder Representation from Transformers, a pre-trained language representation model) algorithm model performs well in understanding text content. The method can use the context-embedded BERT algorithm model to generate dynamic word vectors based on the meaning of words in their specific contexts.

[0053] Optionally, before calculating the standardization of the target information, the method also includes: calculating the content standardization of the target information based on a first Euclidean distance and a second Euclidean distance; the first Euclidean distance is the average Euclidean distance between the word vector at the first position in the target information and the word vector at the corresponding position in other information in the target category in the first cluster; the second Euclidean distance is the average Euclidean distance between the word vector at the second position in the target information and the word vector at the corresponding position in other information in the target category in the second cluster; the word vector at the first position includes the word vector at the first preset number of positions and the word vector at the second preset number of positions, and the word vector at the second position includes the word vector in the word vector sequence except the word vector at the first position.

[0054] In this embodiment, the method can calculate the content standardization of the target information according to the first Euclidean distance and the second Euclidean distance. Specifically, the published content of all information can be clustered according to the difference between the target information and other information in the target category to obtain the first cluster and the second cluster, wherein the first cluster represents the similar parts of all information in the same category, such as the similar parts at the beginning and the end, and the second cluster represents the dissimilar parts of all information in the same category, such as the main part. Then, in the first cluster, the average value of the Euclidean distance between the word vector in the first position in the target information and the word vector in the corresponding position in other information in the same category is calculated, which is recorded as the first Euclidean distance. And in the second cluster, the average value of the Euclidean distance between the word vector in the second position in the target information and the word vector in the corresponding position in other information in the same category is calculated, which is recorded as the second Euclidean distance. Finally, the content standardization of the target information is calculated according to the first Euclidean distance and the second Euclidean distance. The word vectors in corresponding positions are word vectors with the same relative positions in the information in which they are located. For example, for the word vector located first in the target information, the word vector in its corresponding position is the word vector located first in other information; for the word vector located last in the target information, the word vector in its corresponding position is the word vector located last in other information.

[0055] Optionally, before calculating the content standardization of the target information according to the first Euclidean distance and the second Euclidean distance, the method includes: converting all information in the target category into multiple word vector sequences, wherein one piece of information corresponds to one word vector sequence; matching each word vector sequence in the target category with a standard word vector sequence to obtain multiple matching word vectors; clustering the multiple word vector sequences according to the normalized distance of all matching word vectors to obtain a first cluster and a second cluster, wherein the first cluster includes word vectors that are in the first position in each word vector sequence and whose normalized distance of the matching word vectors is less than a preset distance (e.g., 0.7); the second cluster includes word vectors in the multiple word vector sequences except for the word vectors included in the first cluster. It should be noted that the above-mentioned preset distance is only for example and is not limited thereto. In actual applications, the size of the preset distance can be set according to specific needs.

[0056] Before calculating the content standardization of the target information according to the first Euclidean distance and the second Euclidean distance, the method also needs to cluster the published content of all information according to the difference between the target information and other information in the target category. In this embodiment, the method converts all information in the target category into multiple word vector sequences, wherein each piece of information corresponds to a word vector sequence, and matches each word vector sequence in the target category with a standard word vector sequence to obtain multiple matching word vectors, wherein the standard word vector sequence can be any word vector sequence in the target category, but it should be noted that for the same calculation, the same standard word vector sequence is required, that is, each word vector sequence in the target category is matched with the same standard word vector sequence, so that the matching standard can be unified. After obtaining multiple matching word vectors, the multiple word vectors are clustered according to the normalized distance of all matching word vectors to obtain a first cluster and a second cluster, wherein the first cluster includes word vectors that are in the first position in each word vector sequence and the normalized distance of the matching word vector is less than the preset distance; the second cluster includes word vectors in the multiple word vector sequences except the word vectors included in the first cluster.

[0057] In the overall layout of the published content, official information usually adopts some commonly used templates, such as opening remarks, closing remarks, and templates with the same signature, that is, the beginning and end content of the official information may be the same, and the main text content is different. Therefore, this embodiment evaluates the content standardization of the information based on the similarity between the beginning and end content of the information, and the dissimilarity of the main text content. If the similarity between the beginning and end content is very high (or the same), and the similarity between the main text content is low (that is, different), then it can be determined that the content standardization is high, that is, it can be determined that a commonly used template is used, and it is more likely to be official information, and accordingly, its importance should also be higher. Based on the above thinking, all information in the target category can be converted into multiple word vector sequences, where one piece of information corresponds to one word vector sequence. Specifically, the BERT algorithm model can be used to convert all information into multiple word vectors, and then the word vectors are sorted according to the position of the words in the information where they are located to obtain multiple word vector sequences. After obtaining multiple word vector sequences, any word vector sequence is used as a standard word vector sequence, and the DTW (Dynamic Time Warping) algorithm is used for matching to obtain multiple matching word vectors, and the multiple word vector sequences are clustered according to the normalized distance of all matching word vectors.

[0058] The normalized distance of the matching word vectors can represent the similarity between the word vectors with matching positions in the information. The larger the distance, the lower the similarity between the word vectors. Therefore, the word vectors whose normalized distance of the matching word vectors is less than the preset distance can be used as the first cluster, and the first cluster is defined as the similar part; the word vectors whose normalized distance of the matching word vectors is not less than the preset distance can be used as the second cluster, and the second cluster is defined as the dissimilar part. Specifically, for any word vector sequence, the word vectors located at the head and tail parts (the specific number of the head and tail can be determined according to the actual application) can be selected, and the word vectors at the head and tail parts in other information can be traversed and matched to determine whether the normalized distance of the matching word vectors is less than the preset distance. If it is less than the preset distance, the word vector is classified as the first cluster, otherwise it is classified as the second cluster.

[0059] For official information, it is generally concise, and adjectives and adverbs are used less frequently. Therefore, the importance of the information can be calculated based on the proportion of adjectives and adverbs in the information. In this embodiment, the method can use the Jieba algorithm on the published content of the target information, decompose the parts of speech of the text words and sentences in the published content, and then encode them according to the parts of speech, one part of speech corresponds to one encoding, and mark the adjectives and adverbs in the published content and count the number of times the adjectives and adverbs appear, and then further calculate the standardization of the target information based on the similarity between the target information and other information in the target category, the proportion of adjectives and adverbs in the published content of the target information, and the content standardization of the target information.

[0060] Specifically, the target information can be calculated according to the following formula Normative : in, Indicates target information The normativeness of Indicates target information Similarity to other information in the target category to which it belongs; Indicates target information The total number of words in Indicates target information The number of times the pre-set words (i.e. adjectives and adverbs) appear in Indicates target information The proportion of words other than adjectives and adverbs in the target information; Represents the first Euclidean distance, that is, in the first cluster In the target information The word vector at the first position in the target category The average Euclidean distance between word vectors at corresponding positions in a piece of information; Represents the second Euclidean distance, that is, in the second cluster In the target information The word vector at the second position in the target category The average Euclidean distance between word vectors at corresponding positions in a piece of information; Indicates the first The number of the message, Indicates the total amount of information in the target category; Indicates target information The standardization of content.

[0061] In the above formula, the smaller the proportion of adjectives and adverbs in the target information, the more concise the wording of the target information is. The larger the target information is, the Similarity to other information in the target category The weight is used to represent the wording standardization of the target information.

[0062] In the first cluster The higher the similarity between the first and last parts of the target information and the first and last parts of other information (i.e., the first Euclidean distance The smaller the value is, the higher the content standardization of the target information is; in the second cluster The lower the similarity between the middle part of the target information and the middle part of other information (i.e., the second Euclidean distance The larger the value is, the higher the content standardization of the target information is; Indicates target information The standardization of content, The smaller and The larger it is, the more standardized the content of the target information is.

[0063] Optionally, the periodicity of the target information is calculated based on the standardization of the target information and the release time of the target information, including: sorting all the information in the target category according to the release time to obtain the release time series of all the information in the target category; constructing a plane with standardization as the vertical axis and the release time series as the horizontal axis, mapping all the information in the target category to the plane respectively, and dividing them into periods with a preset time length to obtain multiple fluctuation curves, wherein all the information in the same period is connected into one fluctuation curve; calculating the periodicity of the target information based on the standardization of the target information, the standardization of the standard information, and the similarity between the fluctuation curve where the target information is located and the fluctuation curve where the standard information is located, wherein the standard information is information in other periods whose release time matches the release time of the target information.

[0064] For schools, many activities are repeated in a semester or academic year. Therefore, the importance of information can be evaluated by the time regularity of the information release by the information release entity. For example, the higher the probability that information with a strong time regularity is an official release, the higher its importance should be.

[0065] In this embodiment, the method uses a preset time length (such as a natural year) as a cycle, sorts all the information in the target category according to the release time, and obtains the release time sequence of all the information in the target category. Then, a plane with the standard degree as the vertical axis and the release time sequence as the horizontal axis is constructed, and all the information in the target category is mapped to the plane respectively, and is divided with the preset time length as the cycle to obtain multiple fluctuation curves. Among them, all the information in the same cycle is connected into a fluctuation curve. Finally, the periodicity of the target information is calculated based on the standardization of the target information, the standardization of the standard information, and the similarity between the fluctuation curve where the target information is located and the fluctuation curve where the standard information is located, wherein the standard information is the information whose release time in other cycles matches the release time of the target information. The release time matching can be understood as the same position of the release time in the cycle. For example, in the case of a natural year as a cycle, the release time of the information released on August 3 of the previous year matches the release time of the information released on August 3 of the next year.

[0066] Specifically, the target information can be calculated according to the following formula Periodicity : in, Indicates target information The periodicity of Indicates target information The normativeness of represents the total number of standard information corresponding to the target information in the target category, express The standard information in Standard information; Indicates The standardization of the standard information; The fluctuation curve indicating the target information is The similarity of the fluctuation curves of the standard information, Represents an exponential function with the natural constant e as the base.

[0067] In the above formula, Indicates target information With The smaller the value, the greater the similarity between the target information and the information released at the same time point in other periods; The fluctuation curve indicating the target information is The similarity of the fluctuation curve of the standard information. The smaller the value, the stronger the temporal regularity of the information. Indicates target information The comprehensive similarity with all information published in other periods. The smaller the value, the more periodic the target information is. The better.

[0068] Optionally, calculating the importance of the target information based on the periodicity of the target information and the importance of the target category includes: calculating the average value of the periodicity of all information in the target category as the importance of the target category; and calculating the product of the importance of the target category and the periodicity of the target information as the importance of the target information.

[0069] In this embodiment, the method calculates the importance of the target information according to the periodicity of the target information and the importance of the target category, and specifically may include calculating the average value of the periodicity of all information in the target category. as the importance of the target category, and then calculate the importance of the target category Periodicity of information with the target The product of ) as the importance of the target information .

[0070] In this case, the method can further calculate the importance mean of all information in the smart campus community platform , and then according to the importance of the target information The average importance of all information in the smart campus community platform The size of determines the importance of the target information. ≥ , then the target information is considered to be of high importance, and the method stores the target information permanently; conversely, if < , it is considered that the target information is of low importance, and the method performs non-permanent storage of the target information, for example, automatically deleting the target information after six months of storage.

[0071] In this embodiment, the method analyzes the importance of information according to the information release data, and stores information of different importance in different storage methods according to the importance of the information. For example, information of higher importance can be permanently stored, and information of lower importance can be temporarily stored. This can avoid the problem of insufficient storage capacity caused by permanently storing all information, and the inconvenience of deleting important data caused by regularly cleaning up all information. It can also more effectively mine and utilize the value of information, thereby improving campus management efficiency and providing strong support for smart campus management.

[0072] The embodiment of the present invention also provides a smart campus community data storage system, please refer to Figure 3 , the smart campus community data storage system 300 includes: The acquisition module 301 is used to acquire the publishing data of the target information published in the smart campus community platform, wherein the publishing data includes the publishing content and publishing time of the target information; A calculation module 302, configured to calculate the importance of the target information according to the published data; The storage module 303 is used to store the target information according to the importance of the target information, wherein different importance levels correspond to different storage durations.

[0073] Optionally, the published data further includes a publishing subject, and the calculation module 302 calculates the importance of the target information according to the published data, including: Determining the target category to which the target information belongs according to the publisher of the target information; The importance of the target information is calculated according to the publishing content and publishing time of the target information and the publishing data of other information in the target category.

[0074] Optionally, the calculation module 302 calculates the importance of the target information according to the release content and release time of the target information and the release data of other information in the target category, including: Calculate the similarity between the target information and other information in the target category based on the published content of all information in the target category; Calculate the standardization of the target information according to the similarity between the target information and other information in the target category, the proportion of preset words in the published content of the target information, and the standardization of the content of the target information; Calculate the periodicity of the target information according to the standardization of the target information and the release time of the target information; The importance of the target information is calculated according to the periodicity of the target information and the importance of the target category.

[0075] Optionally, the calculation module 302 calculates the similarity between the target information and other information in the target category according to the published content of all information in the target category, including: Numerically processing the published content of all information in the target category to obtain multiple coding curves, wherein one piece of information corresponds to one coding curve, and the target information corresponds to the target coding curve; The similarity between the target information and other information in the target category is calculated according to the length and browsing time of the target coding curve and the length and browsing time of the coding curve of other information in the target category.

[0076] Optionally, the calculation module 302 performs numerical processing on the published content of all information in the target category to obtain multiple coding curves, including: For each piece of information, use text embedding technology to convert the text in the information release content into multiple word vectors; The multiple word vectors are positionally encoded according to the positions of different sentences in the information to obtain corresponding encoding curves.

[0077] Optionally, before calculating the standardization of the target information, the calculation module 302 is further used to: Calculating the content standardization of the target information according to the first Euclidean distance and the second Euclidean distance; The first Euclidean distance is, in the first cluster, an average value of the Euclidean distances between the word vector at the first position in the target information and the word vector at the corresponding position in other information in the target category; The second Euclidean distance is the average value of the Euclidean distances between the word vector at the second position in the target information and the word vector at the corresponding position in other information in the target category in the second cluster; The word vector at the first position includes the word vector located in the front of a first preset number of positions and the word vector located in the back of a second preset number of positions, and the word vector at the second position includes the word vectors in the word vector sequence except the word vector at the first position.

[0078] Optionally, before the calculation module 302 calculates the content standardization of the target information according to the first Euclidean distance and the second Euclidean distance, it is further configured to: Convert all information in the target category into multiple word vector sequences, where one piece of information corresponds to one word vector sequence; Match each word vector sequence in the target category with the standard word vector sequence to obtain multiple matching word vectors; The multiple word vector sequences are clustered according to the normalized distance of all matching word vectors to obtain a first cluster and a second cluster, wherein the first cluster includes word vectors that are in the first position in each word vector sequence and whose normalized distance of matching word vectors is less than a preset distance; and the second cluster includes word vectors in the multiple word vector sequences except the word vectors included in the first cluster.

[0079] Optionally, the calculation module 302 calculates the periodicity of the target information according to the standardization of the target information and the release time of the target information, including: Sorting all the information in the target category according to the release time to obtain a release time sequence of all the information in the target category; Construct a plane with the standardization as the vertical axis and the release time series as the horizontal axis, map all the information in the target category to the plane respectively, and divide it into periods with a preset time length to obtain multiple fluctuation curves, wherein all the information in the same period is connected into one fluctuation curve; The periodicity of the target information is calculated based on the standardization of the target information, the standardization of the standard information, and the similarity between the fluctuation curve of the target information and the fluctuation curve of the standard information, wherein the standard information is information whose release time in other periods matches the release time of the target information.

[0080] Optionally, the calculation module 302 calculates the importance of the target information according to the periodicity of the target information and the importance of the target category, including: Calculate the average value of the periodicity of all information in the target category as the importance of the target category; The product of the importance of the target category and the periodicity of the target information is calculated as the importance of the target information.

[0081] In this embodiment, the smart campus community data storage system analyzes the importance of information based on the information release data, and stores information of different importance in different storage methods according to the importance of the information. For example, information of higher importance can be permanently stored, and information of lower importance can be temporarily stored. This can avoid the problem of insufficient storage capacity caused by permanently storing all information, and the inconvenience of deleting important data caused by regularly cleaning up all information. It can also more effectively mine and utilize the value of information, thereby improving campus management efficiency and providing strong support for smart campus management.

[0082] It should be noted that the sequence of the above embodiments of the present invention is only for description and does not represent the advantages and disadvantages of the embodiments. The processes depicted in the accompanying drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0083] The various embodiments in this specification are described in a progressive manner, and the same or similar parts between the various embodiments can be referenced to each other, and each embodiment focuses on the differences from other embodiments.

Claims

1. A smart campus community data storage method, characterized in that: The method comprises: Acquire the publishing data of the target information published in the smart campus community platform, wherein the publishing data includes the publishing content and publishing time of the target information; Calculating the importance of the target information according to the published data; The target information is stored according to the importance of the target information, wherein different importance levels correspond to different storage durations.

2. The smart campus community data storage method according to claim 1 is characterized in that: The published data also includes a publishing subject, and the calculating the importance of the target information according to the published data includes: Determining the target category to which the target information belongs according to the publisher of the target information; The importance of the target information is calculated according to the publishing content and publishing time of the target information and the publishing data of other information in the target category.

3. The smart campus community data storage method according to claim 2 is characterized in that: The calculating the importance of the target information according to the publishing content and publishing time of the target information and the publishing data of other information in the target category includes: Calculate the similarity between the target information and other information in the target category based on the published content of all information in the target category; Calculate the standardization of the target information according to the similarity between the target information and other information in the target category, the proportion of preset words in the published content of the target information, and the standardization of the content of the target information; Calculate the periodicity of the target information according to the standardization of the target information and the release time of the target information; The importance of the target information is calculated according to the periodicity of the target information and the importance of the target category.

4. The smart campus community data storage method according to claim 3 is characterized in that: The calculating the similarity between the target information and other information in the target category according to the published content of all information in the target category includes: Numerically processing the published content of all information in the target category to obtain multiple coding curves, wherein one piece of information corresponds to one coding curve, and the target information corresponds to the target coding curve; The similarity between the target information and other information in the target category is calculated according to the length and browsing time of the target coding curve and the length and browsing time of the coding curve of other information in the target category.

5. The smart campus community data storage method according to claim 4 is characterized in that: The publication content of all information in the target category is numerically processed to obtain multiple coding curves, including: For each piece of information, use text embedding technology to convert the text in the information release content into multiple word vectors; The multiple word vectors are positionally encoded according to the positions of different sentences in the information to obtain corresponding encoding curves.

6. The smart campus community data storage method according to claim 3 is characterized in that: Before calculating the standardization of the target information, the method further includes: Calculating the content standardization of the target information according to the first Euclidean distance and the second Euclidean distance; The first Euclidean distance is, in the first cluster, an average value of the Euclidean distances between the word vector at the first position in the target information and the word vector at the corresponding position in other information in the target category; The second Euclidean distance is the average value of the Euclidean distances between the word vector at the second position in the target information and the word vector at the corresponding position in other information in the target category in the second cluster; The word vector at the first position includes the word vector located in the front of a first preset number of positions and the word vector located in the back of a second preset number of positions, and the word vector at the second position includes the word vectors in the word vector sequence except the word vector at the first position.

7. The smart campus community data storage method according to claim 6 is characterized in that: Before calculating the content standardization of the target information according to the first Euclidean distance and the second Euclidean distance, the method includes: Convert all information in the target category into multiple word vector sequences, where one piece of information corresponds to one word vector sequence; Match each word vector sequence in the target category with the standard word vector sequence to obtain multiple matching word vectors; The multiple word vector sequences are clustered according to the normalized distance of all matching word vectors to obtain a first cluster and a second cluster, wherein the first cluster includes word vectors that are in the first position in each word vector sequence and whose normalized distance of matching word vectors is less than a preset distance; and the second cluster includes word vectors in the multiple word vector sequences except the word vectors included in the first cluster.

8. The smart campus community data storage method according to claim 3 is characterized in that: The calculating the periodicity of the target information according to the standardization of the target information and the release time of the target information includes: Sorting all the information in the target category according to the release time to obtain a release time sequence of all the information in the target category; Construct a plane with the standardization as the vertical axis and the release time series as the horizontal axis, map all the information in the target category to the plane respectively, and divide it into periods with a preset time length to obtain multiple fluctuation curves, wherein all the information in the same period is connected into one fluctuation curve; The periodicity of the target information is calculated based on the standardization of the target information, the standardization of the standard information, and the similarity between the fluctuation curve of the target information and the fluctuation curve of the standard information, wherein the standard information is information whose release time in other periods matches the release time of the target information.

9. The smart campus community data storage method according to claim 3 is characterized in that: The calculating the importance of the target information according to the periodicity of the target information and the importance of the target category includes: Calculate the average value of the periodicity of all information in the target category as the importance of the target category; The product of the importance of the target category and the periodicity of the target information is calculated as the importance of the target information.

10. A smart campus community data storage system, characterized in that: The system comprises: An acquisition module, used to acquire the publishing data of the target information published in the smart campus community platform, wherein the publishing data includes the publishing content and publishing time of the target information; A calculation module, used for calculating the importance of the target information according to the published data; The storage module is used to store the target information according to the importance of the target information, wherein different importance levels correspond to different storage durations.

Citation Information

Patent Citations

  • System and method for public opinion monitoring analysis and intelligent distribution processing of coordination center

    CN103309948A

  • Method and device for measuring importance of news

    CN104331419A

  • IPFS file storage strategy system based on block chain

    CN114691621A

  • Electronic file backup method and system based on cloud storage

    CN115580623A

  • Low-energy-consumption all-flash storage method and system

    CN117806556A