Spatial-temporal data recommendation method and device, equipment and storage medium

Abstract

The invention relates to the technical field of data processing, and provides a spatio-temporal data recommendation method and device, equipment and a storage medium. The method comprises the following steps: acquiring a retrieval demand and preference information of a target user; performing multi-dimensional retrieval query on the spatio-temporal data according to retrieval requirements to obtain a first data set; the multiple dimensions at least comprise a time dimension, a space dimension, a user dimension, a service type dimension and a function type dimension; based on a preset nearest neighbor service classification model, performing retrieval according to the first data set to obtain a second data set; and performing service recommendation to the target user based on the preference information and the second data set. According to the method, multi-dimensional retrieval is carried out according to the retrieval requirements of the users, expansion and supplement are carried out through the nearest neighbor service classification model, service recommendation is carried out according to user preferences, rapid and effective retrieval of the spatio-temporal data can be achieved under the large-size spatio-temporal data of rural characteristic socialization services, and personalized accurate recommendation is provided for the users.

Description

Technical Field

[0001] The present invention relates to the field of data processing technology, and in particular to a method, device, equipment and storage medium for recommending spatiotemporal data. Background Art

[0002] The data of rural characteristic social services come from a wide range of sources and are of various types. There are significant differences in data structure and semantics. They are multi-source heterogeneous, huge in volume, and highly timely. They are also typical spatiotemporal data. In the existing spatiotemporal data recommendations, a spatial database engine with spatiotemporal indexes is generally used to provide spatial calculation and query interfaces to relational databases, and support a certain degree of spatial information association query. However, for scenarios with large data volumes and strong timeliness, data retrieval mainly relies on manual experience and metadata keywords, and services cannot be accurately matched with needs. There is a lack of query accuracy, and the massive spatiotemporal data of rural characteristic social services cannot meet the needs of fast and accurate queries. Summary of the invention

[0003] The present invention provides a spatiotemporal data recommendation method, device, equipment and storage medium to solve the defect that the existing spatiotemporal data recommendation method has insufficient query accuracy and cannot meet the fast and accurate query requirements in scenarios with large data volume and strong timeliness.

[0004] The present invention provides a spatiotemporal data recommendation method, comprising: Obtaining the search requirements of target users and obtaining the preference information of the target users; Perform a multi-dimensional search query on the spatiotemporal data according to the search requirements to obtain a first data set; the multi-dimensionality includes at least a time dimension, a space dimension, a user dimension, a service type dimension, and a function type dimension; Based on a preset nearest neighbor service classification model, searching according to the first data set to obtain a second data set; A service recommendation is made to the target user based on the preference information and the second data set.

[0005] According to the spatiotemporal data recommendation method provided by the present invention, the step of obtaining the preference information of the target user includes: Acquire the browsing history of the target user, and generate an interest vector of the target user according to the browsing history; Calculating the similarity between the target user and each core user based on the interest vector; the core users are screened from all users based on historical browsing records of all users; If the similarity between the target user and the target core user is greater than a preset threshold, the preference information of the target core user is obtained as the preference information of the target user.

[0006] According to the spatiotemporal data recommendation method provided by the present invention, generating the interest vector of the target user according to the browsing history includes: Determining the browsing time of each preset area by the target user according to the browsing history; Obtaining an initial vector corresponding to each of the preset areas; According to the browsing time, the element value of the first element corresponding to each first area in the initial vector is set to a first eigenvalue, and the element value of the second element corresponding to each second area in the initial vector is set to a second eigenvalue, so as to obtain the interest vector of the target user; the first area is the area in each of the preset areas where the browsing time is greater than the preset time, and the second area is the area other than the first area in each of the preset areas.

[0007] According to the spatiotemporal data recommendation method provided by the present invention, performing a multi-dimensional search query on the spatiotemporal data according to the search requirement to obtain a first data set includes: Determine search keywords according to the search requirements; A multi-dimensional search query is performed on the spatiotemporal data based on the search keywords to obtain a first data set.

[0008] According to the spatiotemporal data recommendation method provided by the present invention, performing a multi-dimensional search query on the target spatiotemporal data set based on the search keyword to obtain a first data set includes: Based on the search keywords, a multi-dimensional search query is performed on the target spatiotemporal data set to obtain the target spatiotemporal data; Performing multi-dimensional data division on the target spatiotemporal data to obtain a plurality of data blocks, and calculating a division coefficient of each of the data blocks during data division; The actual matching accuracy of the multi-dimensional information of the data block is calculated according to the division coefficient; the multi-dimensional information includes at least one of time information, space information, user information, service type information and function information; the multi-dimensional information corresponds to a multi-level index, and the multi-dimensional search query is a link from a node of an upper-level index in the multi-level index to a node of a lower-level index; If the actual matching accuracy is less than the preset ideal matching accuracy, return to and execute the step of performing multi-dimensional data division on the retrieved spatiotemporal data until the actual matching accuracy is greater than or equal to the preset ideal matching accuracy, perform feature fusion on the target spatiotemporal data to obtain a first data set.

[0009] According to the spatiotemporal data recommendation method provided by the present invention, the step of fusing the target spatiotemporal data with features to obtain a first data set includes: Extracting semantic keywords from the target spatiotemporal data to obtain feature segmentations of the target spatiotemporal data; Performing data cleaning and normalization processing on the feature segmentation to obtain standard feature information; Constructing an event tree according to the feature segmentation, and calculating the correlation probability between the standard feature information and the event tree using a multiple correlation coefficient method; Based on the relevant probability, a fuzzy C-means clustering algorithm is used to perform feature fusion on the standard feature information to obtain a first data set.

[0010] According to the spatiotemporal data recommendation method provided by the present invention, the method of retrieving the first data set based on the preset nearest neighbor service classification model to obtain the second data set includes: Acquire the spatiotemporal information of each target spatiotemporal object in the first data set; Using a preset nearest neighbor service classification model, obtaining neighbor objects of each of the spatiotemporal objects through the spatiotemporal information; Retrieving the spatiotemporal data corresponding to the neighboring objects and superimposing them into the first data set to obtain a second data set; The neighboring object is a spatiotemporal object whose spatiotemporal distance to the target spatiotemporal object is less than a preset distance, and the spatiotemporal distance is calculated by the spatiotemporal information of the target spatiotemporal object and the spatiotemporal information of the neighboring object; The nearest neighbor service classification model is obtained by constructing a sample data set based on the spatiotemporal distance between spatiotemporal objects and iteratively training a preset nearest neighbor classification model.

[0011] The present invention also provides a spatiotemporal data recommendation device, comprising the following modules: An information acquisition module is used to acquire the search requirements of target users and obtain the preference information of the target users; A first retrieval module, configured to perform a multi-dimensional retrieval query on the spatiotemporal data according to the retrieval requirements to obtain a first data set; the multi-dimensional retrieval query at least includes a time dimension, a space dimension, a user dimension, a service type dimension, and a function type dimension; A second retrieval module is used to retrieve the first data set based on a preset nearest neighbor service classification model to obtain a second data set; A service recommendation module is used to recommend services to the target user based on the preference information and the second data set.

[0012] The present invention also provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, the steps of any of the above-described spatiotemporal data recommendation methods are implemented.

[0013] The present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, and when the computer program is executed by a processor, the steps of any of the above-mentioned spatiotemporal data recommendation methods are implemented.

[0014] The present invention also provides a computer program product, comprising a computer program, wherein when the computer program is executed by a processor, the steps of any of the above-mentioned spatiotemporal data recommendation methods are implemented.

[0015] The spatiotemporal data recommendation method, device, equipment and storage medium provided by the present invention perform multi-dimensional search and query on spatiotemporal data according to the user's search needs, and on this basis, perform further search and query based on the nearest neighbor service classification model, expand and supplement the spatiotemporal data of the multi-dimensional search and query, and finally perform service recommendation according to the user's preference information, and recommend spatiotemporal data that meets their preferences to the user. Under the large volume of spatiotemporal data of rural characteristic social services, the spatiotemporal data can be quickly and effectively retrieved and personalized and accurately recommended according to the user's search needs, providing users with personalized data services, improving the retrieval timeliness and accuracy of spatiotemporal data, and thus being able to better provide data support for refined management and intelligent services. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are 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.

[0017] Figure 1 It is a flow chart of the spatiotemporal data recommendation method provided by the present invention.

[0018] Figure 2 It is a schematic diagram of the spatiotemporal object classification method provided by the present invention.

[0019] Figure 3 It is a schematic diagram of the spatiotemporal data recommendation process provided by the present invention.

[0020] Figure 4 It is a structural schematic diagram of the spatiotemporal data recommendation device provided by the present invention.

[0021] Figure 5 It is a structural schematic diagram of the electronic device provided by the present invention. DETAILED DESCRIPTION

[0022] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention will be clearly and completely described below in conjunction with the drawings of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0023] The embodiment of the present invention provides a method for recommending spatiotemporal data. Aiming at the accurate and rapid retrieval requirements of spatiotemporal data of rural characteristic social services, according to the characteristics of massive heterogeneous spatiotemporal data of rural characteristic social services, the spatiotemporal data of rural characteristic social services are first classified into types, and semantic feature keywords are extracted on this basis. Standard feature information is obtained after preprocessing, and then a target spatiotemporal data set for query is constructed. Then, a multi-dimensional retrieval query is performed on the spatiotemporal data of rural characteristic social services to complete the rapid matching of information, and finally, spatiotemporal data is recommended according to user interest preferences, so that the recommendation results obtained by users are more accurate.

[0024] Specifically, Figure 1 is a flow chart of the spatiotemporal data recommendation method provided by the present invention, such as Figure 1 As shown, the method comprises the following steps: Step 100, obtaining the search requirements of the target user and obtaining the preference information of the target user; Step 200, performing a multi-dimensional search query on the spatiotemporal data according to the search requirements to obtain a first data set; the multi-dimensionality at least includes a time dimension, a space dimension, a user dimension, a service type dimension, and a function type dimension; Step 300, based on a preset nearest neighbor service classification model, searching according to the first data set to obtain a second data set; Step 400: recommend services to the target user based on the preference information and the second data set.

[0025] First, the search requirements of the target user are obtained, and the preference information of the target user is obtained. Optionally, the search requirements can be configured by the target user in the search interface, and the user's preference information can be generated according to the user's browsing history. The target user can be any user. In this embodiment, the target user is the user who requests spatiotemporal data retrieval.

[0026] According to the search requirements of the target users, a multi-dimensional search query is performed on the target spatiotemporal dataset to obtain a first dataset, wherein the multi-dimensionality includes at least a time dimension, a space dimension, a user dimension, a service type dimension, and a function type dimension. Optionally, when performing a multi-dimensional search query, a search query can be first performed on the target spatiotemporal dataset based on any dimension, and a search query can be performed on the next dimension based on the search results of the dimension until the search query on all dimensions is completed, and the final search query result is the first dataset. Optionally, when performing a multi-dimensional search query, a search query can be performed on the target spatiotemporal dataset based on each dimension, and then the search query results of each dimension are intersected to obtain the first dataset.

[0027] Furthermore, based on the first data set, based on the preset nearest neighbor service classification model, further retrieval is performed based on the first data set to obtain a second data set, which includes the retrieved spatiotemporal data that meets the user's needs, and the neighboring data of the target spatiotemporal data that meets the user's needs.

[0028] Optionally, the spatiotemporal data for user retrieval and query is pre-built and saved in the spatiotemporal database, specifically including the full spatiotemporal data of characteristic rural social services.

[0029] Optionally, when performing multi-dimensional search and query on spatiotemporal data, the search and query dimensions include at least the time dimension, space dimension, user dimension, service type dimension and function type dimension. Furthermore, the time dimension includes different time granularities such as minutes, hours, days, months, quarters and years; the space dimension can be divided according to the scale of administrative divisions, including different spatial scales such as provinces, cities, counties, townships, blocks, plots and service points; the user dimension can be divided according to user type, the service type dimension includes grain, cotton and oil, fruit trees and tea, vegetables and flowers, animal husbandry and fishery, etc., and the function type dimension includes planting, management, production and sales, etc.

[0030] Finally, based on the retrieved second data set, service recommendations are made to the user according to the user's preference information, thereby recommending service categories that meet the user's preferences.

[0031] In this embodiment, a multi-dimensional search query is performed on the spatiotemporal data according to the user's search needs, and on this basis, a further search query is performed based on the nearest neighbor service classification model to supplement the search data. Finally, service recommendations are made based on the user's preference information to recommend spatiotemporal data that meets their preferences to the user. In the case of large-scale spatiotemporal data of rural characteristic social services, fast and effective search and personalized and accurate recommendations can be performed on spatiotemporal data according to the user's search needs, providing users with personalized data services, improving the retrieval timeliness and accuracy of spatiotemporal data, and thus better providing data support for refined management and intelligent services.

[0032] In one embodiment, the user's preference information is generated based on the user's browsing history of spatiotemporal data. In step 100, obtaining the target user's preference information includes: Step 110, obtaining the browsing history of the target user, and generating an interest vector of the target user according to the browsing history; Step 120, calculating the similarity between the target user and each core user based on the interest vector; the core users are screened from the entire user population based on historical browsing records of the entire user population; Step 130: If the similarity between the target user and the target core user is greater than a preset threshold, the preference information of the target core user is obtained as the preference information of the target user.

[0033] When obtaining the preference information of the target user, the browsing history of the target user is first obtained, and the interest vector of the target user is generated according to the browsing history. The interest vector represents the area of ​​interest of the user in the spatiotemporal data. The similarity between the target user and each core user is calculated based on the generated interest vector. If the similarity between the target user and the target core user among the core users is greater than a preset threshold, the preference information of the target core user is obtained as the preference information of the target user, and the preference information of the target user is obtained.

[0034] The preference information of any target user can be generated and updated in real time based on the browsing history of the target user, while the core users are selected from all users based on the historical browsing history of all users and are users who browse spatiotemporal data more frequently.

[0035] Furthermore, in step 110, generating an interest vector of the target user according to the browsing history of the target user specifically includes: Step 111, determining the browsing time of the target user in each preset area according to the browsing record; Step 112, obtaining an initial vector corresponding to each of the preset areas; Step 113, according to the browsing time, the element value of the first element corresponding to each first area in the initial vector is set to the first eigenvalue, and the element value of the second element corresponding to each second area in the initial vector is set to the second eigenvalue, so as to obtain the interest vector of the target user; the first area is the area in each of the preset areas where the browsing time is greater than the preset time, and the second area is the area other than the first area in each of the preset areas.

[0036] According to the browsing history of the target user, the browsing time of the target user in each preset area is determined, and the initial vector corresponding to each preset area is obtained, and the element value in the initial vector is the initial value. According to the browsing time of the target user in each preset area, the element value of the first element corresponding to the first area in the initial vector is set to the first characteristic value, and the element value of the second element corresponding to the second area in the initial vector is set to the second characteristic value, so as to obtain the interest vector of the target user. The element value in the interest vector is used to characterize whether the user is interested in each preset area. The first area is the area in each preset area where the browsing time of the target user is greater than the preset time, and the second area is the area in each preset area except the first area.

[0037] Optionally, the first feature value is different from the second feature value, and illustratively, the first feature value is 1 and the second feature value is 0. The first feature value is used to indicate that the user is interested in the corresponding first area, that is, the browsing time of the user in the first area is greater than a preset time, and the second feature value is used to indicate that the user is not interested in the corresponding second area.

[0038] In one embodiment, based on the area browsed by the user and the length of time the user stays in the area, if the length of time the user stays on a page of a certain spatiotemporal data exceeds a length threshold T, it means that the user is interested in the area.

[0039] For n users and m preset areas, the user set U can be represented as: , ,......, } and the region set R = { , ,......, }, for each user , with corresponding interest vector =( , ,......, ), the interest vectors of all users can constitute n If the user For Region Interested, that is, users If the duration of stay in the area exceeds the time threshold T, then the element value of the corresponding vector =1, if the user Not interested in this area, then the corresponding =0.

[0040] For user similarity, if the user For the target user who requests to recommend spatiotemporal data, user The similarity calculation with other users uses Euclidean Distance, as shown in the following formula: ； (1)

[0041] ； (2)

[0042] Among them, 1≤ ≤n, it can be seen from formula 1 that sim ( , ) value, the more similar the interests of the two users are. , interest vector The proportion of non-zero elements in is the user experience value, which can be expressed as: = ; (3)

[0043] For experience value Greater than the preset threshold (usually set to 0.1) , defined as core users, expressed as .

[0044] According to the interest vectors corresponding to each core user, the interest matrix of the core users can be obtained, which is called the core interest matrix. The user similarity calculation is performed on the basis of the core interest matrix. On the one hand, it can reduce the computational complexity, and on the other hand, it can improve the accuracy of the recommendation. Specifically, due to the large number of users of social service spatiotemporal data, if each user is similar to other users, the amount of calculation is large, the required calculation time is long, and the waiting time of user queries is increased. However, according to user experience, the number of users who need to perform similarity calculation is reduced, and only the similarity comparison with experienced users is required, which reduces the amount of calculation and the calculation time. In addition, the concept of core users is understood as users who browse frequently. Compared with non-core users, core users have rich experience. Taking experienced users as comparison users, the recommendation results of spatiotemporal data are more accurate.

[0045] Optionally, in step 200, a multi-dimensional search query is performed on the target spatiotemporal dataset according to the search requirements of the target user to obtain a first dataset, including: Step 210, determining search keywords according to the search requirements; Step 220: Perform a multi-dimensional search query on the spatiotemporal data based on the search keywords to obtain a first data set.

[0046] A search keyword is determined according to the search requirements of the target user, and the search keyword includes one or more keywords. A multi-dimensional search query is performed on the spatiotemporal data based on the search keyword to obtain a first data set.

[0047] In one embodiment, the retrieved spatiotemporal data is divided into a tree structure, which can classify the spatiotemporal data more quickly and accurately, and prepare for better subsequent integration of the spatiotemporal data. Specifically, step 220 may also include: Step 221, performing a multi-dimensional search query on the target spatiotemporal data set based on the search keyword to obtain the target spatiotemporal data; Step 222, performing multi-dimensional data division on the target spatiotemporal data to obtain a plurality of data blocks, and calculating a division coefficient of each of the data blocks during data division; Step 223, calculating the actual matching accuracy of the multi-dimensional information of the data block according to the partition coefficient; the multi-dimensional information includes at least one of time information, space information, user information, service type information and function information; the multi-dimensional information corresponds to a multi-level index, and the multi-dimensional search query is a link from a node of an upper-level index in the multi-level index to a node of a lower-level index; Step 224, if the actual matching accuracy is less than the preset ideal matching accuracy, return to and execute the step of performing multi-dimensional data division on the retrieved spatiotemporal data until the actual matching accuracy is greater than or equal to the preset ideal matching accuracy, perform feature fusion on the target spatiotemporal data to obtain the first data set.

[0048] After obtaining the target spatiotemporal data through multi-dimensional search and query, the target spatiotemporal data is divided into multiple dimensions to obtain multiple data blocks. Optionally, when multi-dimensional data division is performed on the spatiotemporal data obtained through search and query, multiple data blocks under each dimension can be obtained by dividing the data under the dimension.

[0049] Furthermore, the division coefficient of each data block during data division is calculated, and the actual matching accuracy of the multi-dimensional information of each divided data block is further calculated based on the calculated division coefficient. If the calculated actual matching accuracy is less than the preset ideal matching accuracy, the target spatiotemporal data is re-processed for multi-dimensional data division until the actual matching accuracy is greater than or equal to the preset ideal matching accuracy, and then the target spatiotemporal data is subjected to feature fusion to obtain the first data set. The multi-dimensional information of the data block includes at least one of time information, space information, user information, service type information, and function information; the multi-dimensional information corresponds to a multi-level index, and the multi-dimensional search query is a link from the node of the upper index in the multi-level index to the node of the lower index. That is, the multi-dimensional search query of the spatiotemporal data is to perform a search query of the next dimension on the basis of the result of the search query of the previous dimension, until the search query of all dimensions is completed to obtain the target spatiotemporal data.

[0050] Optionally, the actual matching accuracy of the multi-dimensional information of each data block is calculated, specifically, the matching rate between the multi-dimensional information of the data block and the multi-dimensional information when the data is divided is calculated.

[0051] Optionally, the division coefficient is calculated as shown in the following formula 4: ; (4)

[0052] in, Indicates the division coefficient when dividing spatiotemporal data; Represents the partition error factor, and its value range is generally (0.1, 0.2); N Indicates the number of data blocks obtained by dividing the spatiotemporal data; i express N The first i data blocks; Indicates i The amount of data contained in each data block; Represents the total amount of target spatiotemporal data.

[0053] Furthermore, based on Formula 4, the actual matching accuracy of multi-dimensional information such as time information, space information, user information, service type information and function information of spatiotemporal data is calculated in the manner shown in the following Formula 5: ; (5)

[0054] In formula 5, It represents the actual matching accuracy of multi-dimensional information such as time information, space information, user information, service type information and function information of spatiotemporal data, and its value range is generally (0, 1); It represents the partition coefficient when partitioning the target spatiotemporal data, also known as the matching coefficient, and its value range is generally (0.6, 0.8); Indicates the actual matching degree when the time information, location information and content information of spatiotemporal data match; Indicates the ideal matching degree when the time information, location information and content information of spatiotemporal data match; Represents the matching error coefficient, and its value range is generally (0.15, 0.2).

[0055] The actual matching accuracy calculated The target spatiotemporal data are compared with the preset ideal matching accuracy. If the calculated actual matching accuracy is less than the ideal matching accuracy, it is determined that the multi-dimensional information matching of the time information, space information, user information, service type information and function information of the target spatiotemporal data is unqualified, and the target spatiotemporal data are re-divided into multi-dimensional data according to the time dimension, space dimension, user dimension, service type dimension and function dimension until the calculated actual matching accuracy is greater than or equal to the ideal matching accuracy. The target spatiotemporal data obtained by the multi-dimensional retrieval query are feature fused to obtain the first data set.

[0056] By performing multi-dimensional data partitioning processing on spatiotemporal data in a tree structure, spatiotemporal data can be classified more quickly and accurately, thus facilitating subsequent feature fusion.

[0057] Further, after the retrieved target spatiotemporal data is divided, a first data set is obtained by feature fusion. Based on this, in step 224, feature fusion is performed on the target spatiotemporal data to obtain the first data set, which further includes: Step 2241, extracting semantic keywords from the target spatiotemporal data to obtain feature segmentations of the target spatiotemporal data; Step 2242, performing data cleaning and normalization processing on the feature segmentation to obtain standard feature information; Step 2243, constructing an event tree according to the feature segmentation, and calculating the correlation probability between the standard feature information and the event tree using a multiple correlation coefficient method; Step 2244: Based on the relevant probability, a fuzzy C-means clustering algorithm is used to perform feature fusion on the standard feature information to obtain a first data set.

[0058] When performing feature fusion, firstly, semantic keywords are extracted from the acquired target spatiotemporal data to obtain the feature segmentation of the target spatiotemporal data, and the extracted feature segmentation is cleaned and normalized to obtain standard feature information. An event tree is constructed based on the extracted feature segmentation, and the correlation probability between the standard feature information and the event tree is calculated using the multiple correlation coefficient method. Finally, based on the calculated correlation probability, the fuzzy C-means clustering algorithm is used to perform feature fusion on the standard feature information obtained by preprocessing to obtain the first data set.

[0059] Further, based on the preset nearest neighbor service classification model, further retrieval is performed according to the retrieved first data set to obtain the second data set, which is achieved by classifying and retrieving the spatiotemporal objects in the spatiotemporal data based on the constructed nearest neighbor service classification model. Specifically, step 300 specifically includes: Step 310, obtaining the spatiotemporal information of each target spatiotemporal object in the first data set; Step 320, using a preset nearest neighbor service classification model to obtain neighbor objects of each of the spatiotemporal objects through the spatiotemporal information; Step 330, retrieving the spatiotemporal data corresponding to the neighboring objects and superimposing them into the first data set to obtain a second data set; The neighboring object is a spatiotemporal object whose spatiotemporal distance to the target spatiotemporal object is less than a preset distance, and the spatiotemporal distance is calculated by the spatiotemporal information of the target spatiotemporal object and the spatiotemporal information of the neighboring object; The nearest neighbor service classification model is obtained by constructing a sample data set based on the spatiotemporal distance between spatiotemporal objects and iteratively training a preset nearest neighbor classification model.

[0060] First, the spatiotemporal information of each target spatiotemporal object in the first data set is obtained, and the neighboring objects of each target spatiotemporal object are obtained through the spatiotemporal information of each target spatiotemporal object using a preset nearest neighbor service classification model. Optionally, any target spatiotemporal object has K neighboring objects, where K is a positive integer.

[0061] The spatiotemporal data corresponding to each neighboring object is retrieved and superimposed on the first data set to obtain a second data set. Optionally, the neighboring object is a spatiotemporal object whose spatiotemporal distance from the target spatiotemporal object is less than a preset distance, and the spatiotemporal distance between the neighboring object and the target spatiotemporal object is calculated based on the spatiotemporal information of the target spatiotemporal object and the spatiotemporal information of the neighboring spatiotemporal object. Based on this, the retrieved second data set contains the target spatiotemporal data that meets the user's needs, as well as the adjacent spatiotemporal data of the target spatiotemporal data.

[0062] Optionally, the spatiotemporal data stored in the spatiotemporal database can be Figure 2The space-time object shown is constructed in the form of a space-time object, and each space-time object has K neighboring objects within a preset space-time distance range. Based on this, when the space-time information of any space-time object is obtained, the K neighboring objects of the space-time object within the preset space-time distance range can be quickly matched according to the space-time information, and the space-time data corresponding to the K neighboring objects can be retrieved. Optionally, when the preset space-time distance range is the same, different space-time objects can have different K values; or, when different space-time objects have the same K value, the preset space-time distance ranges corresponding to different space-time objects can be different, and there is no specific limitation on this.

[0063] Figure 2 The classification method of some spatiotemporal objects is given as an example. Different spatiotemporal objects are divided from the time dimension and the space dimension according to their spatiotemporal information. From the time dimension, different time points such as time 1 to time 5 are divided, and from the space dimension, different location areas such as area 1 to area 5 are divided. Each spatiotemporal object is matched according to the time point and location area corresponding to its spatiotemporal information, and the matching degree between the multi-dimensional information of the spatiotemporal object and the divided time 1 to time 5 and area 1 to area 5 is calculated, and then the actual matching accuracy of the data block is calculated in accordance with the method shown in Formula 5 in combination with the division coefficient.

[0064] Furthermore, the nearest neighbor service classification model is obtained by iteratively training the preset nearest neighbor classification model based on a sample data set constructed based on the spatiotemporal distance between spatiotemporal objects. That is, when constructing the nearest neighbor service classification model, the spatiotemporal information of each spatiotemporal object is first determined, and the spatiotemporal information includes at least time information and space information. The spatiotemporal distance between spatiotemporal objects is calculated based on the spatiotemporal information. The spatiotemporal distance between spatiotemporal objects is used as the basis for model optimization, and a sample data set is constructed. The preset nearest neighbor classification model is iteratively trained to obtain the nearest neighbor service classification model. Among them, the spatiotemporal object corresponds to the service category, and the spatiotemporal distance between the spatiotemporal objects is also the distance between the service categories. Any service category has K (K is a positive integer) service category neighbors within the preset spatiotemporal distance range, which are called neighbor objects.

[0065] Optionally, the spatiotemporal distance between spatiotemporal objects is measured and calculated using Manhattan distance and Euclidean distance as shown in the following formula 6-7: ( , ) ; (6)

[0066] ( , ) , ( , ) ; (7)

[0067] In formula 6-7, n represents the number of users. Indicates l User's i A space-time object, Indicates l User's j A space-time object, ( , ) indicates the i The space-time object and j The space-time distance between two space-time objects. When p=1, it represents the Manhattan distance, and when p=2, it represents the Euclidean distance.

[0068] In one embodiment, referring to Figure 3 The spatiotemporal data recommendation process shown in the figure, the spatiotemporal data recommendation method provided by the embodiment of the present invention mainly includes the stages of spatiotemporal data division, multi-dimensional matching of spatiotemporal data, fusion of spatiotemporal data features, and personalized recommendation of spatiotemporal data.

[0069] In the spatiotemporal data division stage, a multi-dimensional search query is performed according to the user's search requirements to obtain the target spatiotemporal data, and the target spatiotemporal data is subjected to multi-dimensional data division processing to obtain multiple corresponding data blocks, and the division coefficient of each data block is calculated in the manner shown in Formula 4. In the multi-dimensional matching stage of spatiotemporal data, the actual matching accuracy of the multi-dimensional information of each data block is calculated in the manner shown in Formula 5. If the actual matching accuracy is less than the preset ideal matching accuracy, the target spatiotemporal data is re-divided into multi-dimensional processing until the actual matching accuracy is greater than or equal to the preset ideal matching accuracy.

[0070] Optionally, a tree structure is used to perform multi-dimensional data partitioning processing on the target spatiotemporal data to facilitate subsequent rapid query and feature fusion of the spatiotemporal data to obtain a first data set.

[0071] Furthermore, in the spatiotemporal data feature fusion stage, semantic feature keywords are extracted from the target spatiotemporal data to obtain the feature segmentation of the target spatiotemporal data and preprocess it to obtain standard feature information, and an event tree is constructed based on the extracted feature segmentation. The preprocessing of the feature segmentation includes at least data cleaning and normalization.

[0072] Furthermore, in the personalized recommendation stage, based on the preset nearest neighbor service classification model, the service category is used as the spatiotemporal object, and the spatiotemporal data is further searched and queried according to the first data set to obtain the second data set. Specifically, first, according to the spatiotemporal information of each target spatiotemporal object in the first data set, the neighboring objects of each target spatiotemporal object are obtained, and then the spatiotemporal data corresponding to the neighboring objects are retrieved and superimposed on the first data set to supplement the first data set. Finally, according to the user preference information, the service category that matches the user preference is selected from the second data set to recommend services to the user.

[0073] Optionally, the target user's preference information is based on the similarity between the target user and the core user. According to the preference information of the core user, there is a certain difference in the similarity between the target user and the core user. The data supplement of the first data set based on the nearest neighbor service classification model can be combined with the core user's experience data based on the user's retrieval needs to infer the user's area of ​​interest, thereby improving the recommendation accuracy of spatiotemporal data.

[0074] In this embodiment, based on the precise and rapid retrieval requirements of large-scale spatiotemporal data of rural characteristic social services, personalized recommendations for spatiotemporal data are achieved through multi-dimensional retrieval queries and semantic keyword extraction of spatiotemporal data, and extended queries can be performed according to retrieval requirements and the nearest neighbor service classification model, and potential user-interested data can be inferred in combination with user preference information. Compared with the manual search for recommended data sets based on experience, the recommendation results are more scientific and accurate, which improves the utilization rate of massive spatiotemporal data of rural characteristic social services and provides accurate and reliable data support for the refined management and intelligent services of rural characteristic products.

[0075] Furthermore, based on the user's retrieval needs, the nearest neighbor service classification model is used to expand and supplement the user's retrieval needs. When obtaining user preference information, the experience of core users is combined. By calculating the similarity between the target user and the core user, the target user's area of ​​interest is inferred. This not only reduces the amount of calculation when obtaining user preferences, but also improves the accuracy of service recommendations.

[0076] The spatiotemporal data recommendation device provided by the present invention is described below. The spatiotemporal data recommendation device described below and the spatiotemporal data recommendation method described above can be referenced to each other.

[0077] Reference Figure 4 , the spatiotemporal data recommendation device provided by the embodiment of the present invention includes: The information acquisition module 10 is used to acquire the search requirements of the target user and the preference information of the target user; A first retrieval module 20, configured to perform a multi-dimensional retrieval query on the spatiotemporal data according to the retrieval requirements to obtain a first data set; the multi-dimensional retrieval query at least includes a time dimension, a space dimension, a user dimension, a service type dimension, and a function type dimension; A second retrieval module 30, configured to retrieve the first data set based on a preset nearest neighbor service classification model to obtain a second data set; The service recommendation module 40 is used to recommend services to the target user based on the preference information and the second data set.

[0078] In one embodiment, the information acquisition module 10 is further used to: Acquire the browsing history of the target user, and generate an interest vector of the target user according to the browsing history; Calculating the similarity between the target user and each core user based on the interest vector; the core users are screened from all users based on historical browsing records of all users; If the similarity between the target user and the target core user is greater than a preset threshold, the preference information of the target core user is obtained as the preference information of the target user.

[0079] In one embodiment, the information acquisition module 10 is further used to: Determining the browsing time of each preset area by the target user according to the browsing history; Obtaining an initial vector corresponding to each of the preset areas; According to the browsing time, the element value of the first element corresponding to each first area in the initial vector is set to a first eigenvalue, and the element value of the second element corresponding to each second area in the initial vector is set to a second eigenvalue, so as to obtain the interest vector of the target user; the first area is the area in each of the preset areas where the browsing time is greater than the preset time, and the second area is the area other than the first area in each of the preset areas.

[0080] In one embodiment, the first retrieval module 20 is further configured to: Determine search keywords according to the search requirements; A multi-dimensional search query is performed on the spatiotemporal data based on the search keywords to obtain a first data set.

[0081] In one embodiment, the first retrieval module 20 is further configured to: Based on the search keywords, a multi-dimensional search query is performed on the target spatiotemporal data set to obtain the target spatiotemporal data; Performing multi-dimensional data division on the target spatiotemporal data to obtain a plurality of data blocks, and calculating a division coefficient of each of the data blocks during data division; The actual matching accuracy of the multi-dimensional information of the data block is calculated according to the division coefficient; the multi-dimensional information includes at least one of time information, space information, user information, service type information and function information; the multi-dimensional information corresponds to a multi-level index, and the multi-dimensional search query is a link from a node of an upper-level index in the multi-level index to a node of a lower-level index; If the actual matching accuracy is less than the preset ideal matching accuracy, return to and execute the step of performing multi-dimensional data division on the retrieved spatiotemporal data until the actual matching accuracy is greater than or equal to the preset ideal matching accuracy, perform feature fusion on the target spatiotemporal data to obtain a first data set.

[0082] In one embodiment, the first retrieval module 20 is further configured to: Extracting semantic keywords from the target spatiotemporal data to obtain feature segmentations of the target spatiotemporal data; Performing data cleaning and normalization processing on the feature segmentation to obtain standard feature information; Constructing an event tree according to the feature segmentation, and calculating the correlation probability between the standard feature information and the event tree using a multiple correlation coefficient method; Based on the relevant probability, a fuzzy C-means clustering algorithm is used to perform feature fusion on the standard feature information to obtain a first data set.

[0083] In one embodiment, the second retrieval module 30 is further configured to: Acquire the spatiotemporal information of each target spatiotemporal object in the first data set; Using a preset nearest neighbor service classification model, obtaining neighbor objects of each of the spatiotemporal objects through the spatiotemporal information; Retrieving the spatiotemporal data corresponding to the neighboring objects and superimposing them into the first data set to obtain a second data set; The neighboring object is a spatiotemporal object whose spatiotemporal distance to the target spatiotemporal object is less than a preset distance, and the spatiotemporal distance is calculated by the spatiotemporal information of the target spatiotemporal object and the spatiotemporal information of the neighboring object; The nearest neighbor service classification model is obtained by constructing a sample data set based on the spatiotemporal distance between spatiotemporal objects and iteratively training a preset nearest neighbor classification model.

[0084] Figure 5 An example of a physical structure diagram of an electronic device is shown in FIG. Figure 5As shown, the electronic device may include: a processor 510, a communications interface 520, a memory 530 and a communication bus 540, wherein the processor 510, the communications interface 520 and the memory 530 communicate with each other through the communication bus 540. The processor 510 may call the logic instructions in the memory 530 to execute the steps of the spatiotemporal data recommendation method, for example including: Obtaining the search requirements of target users and obtaining the preference information of the target users; Perform a multi-dimensional search query on the spatiotemporal data according to the search requirements to obtain a first data set; the multi-dimensionality includes at least a time dimension, a space dimension, a user dimension, a service type dimension, and a function type dimension; Based on a preset nearest neighbor service classification model, searching according to the first data set to obtain a second data set; A service recommendation is made to the target user based on the preference information and the second data set.

[0085] In addition, the logic instructions in the above-mentioned memory 530 can be implemented in the form of a software functional unit and can be stored in a computer-readable storage medium when it is sold or used as an independent product. Based on such an understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art or the part of the technical solution, can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, etc. Various media that can store program codes.

[0086] On the other hand, the present invention further provides a computer program product, the computer program product includes a computer program, the computer program can be stored in a non-transitory computer-readable storage medium, and when the computer program is executed by a processor, the computer can perform the steps of the spatiotemporal data recommendation method provided in the above embodiments, for example, including: Obtaining the search requirements of target users and obtaining the preference information of the target users; Perform a multi-dimensional search query on the spatiotemporal data according to the search requirements to obtain a first data set; the multi-dimensionality includes at least a time dimension, a space dimension, a user dimension, a service type dimension, and a function type dimension; Based on a preset nearest neighbor service classification model, searching according to the first data set to obtain a second data set; A service recommendation is made to the target user based on the preference information and the second data set.

[0087] In another aspect, the present invention further provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, is implemented to perform the steps of the spatiotemporal data recommendation method provided in the above embodiments, for example, including: Obtaining the search requirements of target users and obtaining the preference information of the target users; Perform a multi-dimensional search query on the spatiotemporal data according to the search requirements to obtain a first data set; the multi-dimensionality includes at least a time dimension, a space dimension, a user dimension, a service type dimension, and a function type dimension; Based on a preset nearest neighbor service classification model, searching according to the first data set to obtain a second data set; A service recommendation is made to the target user based on the preference information and the second data set.

[0088] The device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Ordinary technicians in this field can understand and implement it without paying creative labor.

[0089] Through the description of the above implementation methods, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a necessary general hardware platform, and of course, can also be implemented by hardware. Based on this understanding, the above technical solution is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0090] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A spatiotemporal data recommendation method, characterized in that: include: Obtaining the search requirements of target users and obtaining the preference information of the target users; Perform a multi-dimensional search query on the spatiotemporal data according to the search requirements to obtain a first data set; the multi-dimensionality includes at least a time dimension, a space dimension, a user dimension, a service type dimension, and a function type dimension; Based on a preset nearest neighbor service classification model, searching according to the first data set to obtain a second data set; A service recommendation is made to the target user based on the preference information and the second data set.

2. The spatiotemporal data recommendation method according to claim 1, characterized in that: The obtaining the preference information of the target user includes: Acquire the browsing history of the target user, and generate an interest vector of the target user according to the browsing history; Calculating the similarity between the target user and each core user based on the interest vector; the core users are screened from all users based on historical browsing records of all users; If the similarity between the target user and the target core user is greater than a preset threshold, the preference information of the target core user is obtained as the preference information of the target user.

3. The spatiotemporal data recommendation method according to claim 2, characterized in that: The generating the interest vector of the target user according to the browsing history includes: Determining the browsing time of each preset area by the target user according to the browsing history; Obtaining an initial vector corresponding to each of the preset areas; According to the browsing time, the element value of the first element corresponding to each first area in the initial vector is set to a first eigenvalue, and the element value of the second element corresponding to each second area in the initial vector is set to a second eigenvalue, so as to obtain the interest vector of the target user; the first area is the area in each of the preset areas where the browsing time is greater than the preset time, and the second area is the area other than the first area in each of the preset areas.

4. The spatiotemporal data recommendation method according to claim 1, characterized in that: The performing a multi-dimensional search query on the spatiotemporal data according to the search requirement to obtain a first data set includes: Determine search keywords according to the search requirements; A multi-dimensional search query is performed on the spatiotemporal data based on the search keywords to obtain a first data set.

5. The spatiotemporal data recommendation method according to claim 4, characterized in that: The step of performing a multi-dimensional search query on the target spatiotemporal dataset based on the search keyword to obtain a first dataset includes: Based on the search keywords, a multi-dimensional search query is performed on the target spatiotemporal data set to obtain the target spatiotemporal data; Performing multi-dimensional data division on the target spatiotemporal data to obtain a plurality of data blocks, and calculating a division coefficient of each of the data blocks during data division; The actual matching accuracy of the multi-dimensional information of the data block is calculated according to the division coefficient; the multi-dimensional information includes at least one of time information, space information, user information, service type information and function information; the multi-dimensional information corresponds to a multi-level index, and the multi-dimensional search query is a link from a node of an upper-level index in the multi-level index to a node of a lower-level index; If the actual matching accuracy is less than the preset ideal matching accuracy, return to and execute the step of performing multi-dimensional data division on the retrieved spatiotemporal data until the actual matching accuracy is greater than or equal to the preset ideal matching accuracy, perform feature fusion on the target spatiotemporal data to obtain a first data set.

6. The method for recommending spatiotemporal data according to claim 5, characterized in that: The step of performing feature fusion on the target spatiotemporal data to obtain a first data set includes: Extracting semantic keywords from the target spatiotemporal data to obtain feature segmentations of the target spatiotemporal data; Performing data cleaning and normalization processing on the feature segmentation to obtain standard feature information; Constructing an event tree according to the feature segmentation, and calculating the correlation probability between the standard feature information and the event tree using a multiple correlation coefficient method; Based on the relevant probability, a fuzzy C-means clustering algorithm is used to perform feature fusion on the standard feature information to obtain a first data set.

7. The spatiotemporal data recommendation method according to claim 1, characterized in that: The method of performing a search based on the preset nearest neighbor service classification model according to the first data set to obtain a second data set includes: Acquire the spatiotemporal information of each target spatiotemporal object in the first data set; Using a preset nearest neighbor service classification model, obtaining neighbor objects of each of the spatiotemporal objects through the spatiotemporal information; Retrieving the spatiotemporal data corresponding to the neighboring objects and superimposing them into the first data set to obtain a second data set; The neighboring object is a spatiotemporal object whose spatiotemporal distance to the target spatiotemporal object is less than a preset distance, and the spatiotemporal distance is calculated by the spatiotemporal information of the target spatiotemporal object and the spatiotemporal information of the neighboring object; The nearest neighbor service classification model is obtained by constructing a sample data set based on the spatiotemporal distance between spatiotemporal objects and iteratively training a preset nearest neighbor classification model.

8. A spatiotemporal data recommendation device, characterized in that: include: An information acquisition module is used to acquire the search requirements of target users and obtain the preference information of the target users; A first retrieval module, configured to perform a multi-dimensional retrieval query on the spatiotemporal data according to the retrieval requirements to obtain a first data set; the multi-dimensional retrieval query at least includes a time dimension, a space dimension, a user dimension, a service type dimension, and a function type dimension; A second retrieval module is used to retrieve the first data set based on a preset nearest neighbor service classification model to obtain a second data set; A service recommendation module is used to recommend services to the target user based on the preference information and the second data set.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the computer program, the steps of the spatiotemporal data recommendation method according to any one of claims 1 to 7 are implemented.

10. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the spatiotemporal data recommendation method according to any one of claims 1 to 7 are implemented.

Images