Point of interest processing method and apparatus, electronic device, medium, and program product

Abstract

The present disclosure provides a point of interest processing method and device, electronic equipment, medium and program product, relates to the technical field of artificial intelligence, in particular to the technical field of map. The method comprises the following steps: obtaining a fused point of interest, the fused point of interest being obtained by data fusion on information of at least one original point of interest; determining whether the fused point of interest meets an online condition; if not, determining an update factor of the fused point of interest, and when the update factor is greater than a first preset value, obtaining a supplementary point of interest corresponding to the fused point of interest according to the information of the fused point of interest, and updating the fused point of interest according to the supplementary point of interest. The method improves the accuracy of the point of interest.

Description

TECHNICAL FIELD

[0001] The present disclosure relates to the map technology in the field of artificial intelligence technology, in particular to a point of interest (POI) processing method and device, electronic equipment, medium and program product. BACKGROUND

[0002] A point of interest is a point that can attract attention in a map, such as a scenic spot, a company, a shopping mall, a store, a bus stop, etc. The information of a point of interest generally includes name, address, phone number, business description, classification, and on-site photos.

[0003] Since various new or changed locations may appear in the real world, it is necessary to add or update the points of interest in the map accordingly to ensure the accuracy of the map. When obtaining the information of a point of interest, the point of interest can be extracted from data such as WIFI information, logistics information, street view pictures, Internet information, or user reported information. However, in these methods, the quality of the obtained data cannot be guaranteed, which often leads to inaccurate points of interest. SUMMARY

[0004] The present disclosure provides a point of interest processing method and device, electronic equipment, medium and program product that improve the accuracy of points of interest.

[0005] According to an aspect of the present disclosure, a point of interest processing method is provided, comprising:

[0006] obtaining a fused point of interest, the fused point of interest being obtained by data fusion on information of at least one original point of interest;

[0007] determining whether the fused point of interest meets an online condition;

[0008] if not, determining an update factor of the fused point of interest, and when the update factor is greater than a first preset value, obtaining a supplementary point of interest corresponding to the fused point of interest according to the information of the fused point of interest, to update the fused point of interest according to the supplementary point of interest.

[0009] According to another aspect of the present disclosure, a point of interest processing device is provided, comprising:

[0010] an obtaining module configured to obtain a fused point of interest, the fused point of interest being obtained by data fusion on information of at least one original point of interest;

[0011] a determining module configured to determine whether the fused point of interest meets an online condition;

[0012] The updating module is configured to determine an updating factor of the fused interest point if the condition is not met, and acquire a supplementary interest point corresponding to the fused interest point according to information of the fused interest point if the updating factor is greater than a first preset value, so as to update the fused interest point according to the supplementary interest point.

[0013] According to still another aspect of the present disclosure, an electronic device is provided, comprising:

[0014] at least one processor; and

[0015] a memory connected with the at least one processor in communication; wherein

[0016] The memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the method of the first aspect.

[0017] According to still another aspect of the present disclosure, a non-transitory computer readable storage medium storing computer instructions is provided, the computer instructions being used to cause a computer to perform the method of the first aspect.

[0018] According to still another aspect of the present disclosure, a computer program product is provided, the program product comprising: a computer program stored in a readable storage medium, the computer program being readable by at least one processor of an electronic device, the at least one processor executing the computer program to cause the electronic device to perform the method of the first aspect.

[0019] According to the technical solution of the present disclosure, the accuracy of interest points is improved.

[0020] It should be understood that the content described in this part is not intended to identify key or important features of the embodiments of the present disclosure, nor to limit the scope of the present disclosure. Other features of the present disclosure will become apparent from the following description. BRIEF DESCRIPTION OF DRAWINGS

[0021] The accompanying drawings are used to better understand the present solution and do not limit the present disclosure. Among them:

[0022] Figure 1 is a schematic diagram of an interest point in a map;

[0023] Figure 2 is a flowchart of a processing method of an interest point according to an embodiment of the present disclosure;

[0024] Figure 3 is a structural schematic diagram of a processing device of an interest point according to an embodiment of the present disclosure;

[0025] Figure 4 This is a schematic block diagram of an electronic device used to implement the point of interest processing method of the embodiments of this disclosure. Detailed Implementation

[0026] The exemplary embodiments of this disclosure are described below with reference to the accompanying drawings, including various details of the embodiments to aid understanding, and should be considered merely exemplary. Therefore, those skilled in the art will recognize that various changes and modifications can be made to the embodiments described herein without departing from the scope and spirit of this disclosure. Similarly, for clarity and brevity, descriptions of well-known functions and structures are omitted in the following description.

[0027] Figure 1 It is a schematic diagram of points of interest in a map, such as... Figure 1 As shown in the image, the map displays various locations, each of which is a point of interest (POI). First, we introduce the methods used in related technologies to acquire new POIs through different data collection approaches.

[0028] For example, obtaining Wi-Fi information can yield the Wi-Fi name and the user's location when connecting to the Wi-Fi. For instance, in 'yiyihuoguo_5G', 'yiyihuoguo' can be identified as a newly added point of interest. However, we can also see that 'yiyihuoguo' is a name using pinyin, and the correct Chinese name cannot be determined from this information. Since most Wi-Fi names are personalized by the user and are often not standardized, this method often cannot accurately obtain the name of the point of interest.

[0029] For example, when obtaining logistics information, such as the address "XX Hair Salon, No. 11, First Street, Cixi City, Ningbo, Zhejiang Province", it can be identified that "XX Hair Salon" may be a new point of interest. However, the logistics information only contains address information, and the coordinate information often needs to be generated based on the address, resulting in relatively low coordinate accuracy.

[0030] For example, street view images are obtained from map field data collection vehicles. Image recognition or text recognition is then performed to obtain points of interest (POIs). However, due to the presence of many obstructions on the road, such as trees and signs, POIs in the street view images are often blocked, leading to inaccurate identification of POIs.

[0031] For example, users upload information about new points of interest (POIs) through the corresponding entry point on a map app, and map professionals then process the POIs accordingly. However, the number of users who upload information is small, the accuracy of the information uploaded by users cannot be guaranteed, and the scenarios reported by users often mean that users search for the POI on the map but cannot find it, and their search needs are not being met.

[0032] For example, the information of the interest point is obtained based on the Internet, but the accuracy of the information obtained in this way is also uncontrollable.

[0033] As can be seen from the above introduction, the information obtained by a single data collection method often has low data quality, which leads to inaccurate interest points determined by the information. Therefore, in the embodiments of the present disclosure, the interest points determined by different methods (different sources) are taken as original interest points, and the original interest points are fused to obtain fused interest points, so as to improve the accuracy of the interest points. After obtaining the fused interest points, it is further judged whether the fused interest points can be online. For the fused interest points that do not meet the online conditions, data is further collected to obtain corresponding supplementary original interest points, and the supplementary original interest points and the fused interest points are further fused, so as to improve the accuracy of the fused interest points.

[0034] The present disclosure provides a processing method and device of an interest point, an electronic device, a medium and a program product, which are applied to the field of map technology in the field of artificial intelligence technology, and can be applied to scenarios such as map updating to improve the accuracy of the interest point.

[0035] In the technical solution of the present disclosure, the collection, storage, use, processing, transmission, provision and disclosure of user personal information comply with relevant laws and regulations and do not violate public order and good customs.

[0036] In the following, the processing method of the interest point provided by the present disclosure will be described in detail through specific embodiments. It can be understood that the following specific embodiments can be combined with each other, and the same or similar concepts or processes may not be described in detail in some embodiments.

[0037] Figure 2 is a flowchart of a processing method of an interest point according to an embodiment of the present disclosure. The execution subject of the method is a processing device of an interest point, which can be realized by software and / or hardware. As shown in Figure 2 the method comprises:

[0038] S201, obtaining a fused interest point, the fused interest point being obtained by data fusion on information of at least one original interest point.

[0039] The original interest point in the embodiments of the present disclosure refers to an interest point determined by data obtained through a data collection manner (or referred to as a data source), for example, an interest point determined by collecting street view pictures taken by a car is an original interest point, an interest point determined by WIFI information is an original interest point, and an interest point determined by logistics information is an original interest point, and the fused interest point is obtained by fusing at least one original interest point. For example, for a store, the original interest points corresponding to the store are determined through WIFI information and logistics information respectively, the two original interest points are fused to obtain the fused interest point corresponding to the store, and the accuracy of the fused interest point obtained by fusing the original interest points is higher. It can be understood that if there is only one original interest point at a certain location, the original interest point can also be used as a fused interest point.

[0040] In S202, it is determined whether the fused interest point meets the online condition, if yes, S203 is performed, and if no, S204 is performed.

[0041] In S203, the fused interest point is put online.

[0042] In S204, an update factor of the fused interest point is determined, and if the update factor is greater than a first preset value, a supplementary interest point corresponding to the fused interest point is obtained according to the information of the fused interest point, so as to update the fused interest point according to the supplementary interest point.

[0043] For the fused interest point, the accuracy is improved compared with the original interest point, but if the information of the original interest point corresponding to the fused interest point is not accurate, the information of the fused interest point obtained after fusion may still be not accurate enough, and therefore, it is further determined whether the fused interest point can be put online, and the fused interest point can be put online only when the online condition is met.

[0044] For the fused interest point that does not meet the online condition, it is further determined the update factor of the fused interest point, the update factor is used to represent the value or importance of the fused interest point, if the update factor of the fused interest point is greater than a first preset value, it means that the fused interest point has high value or is important, and therefore, a supplementary interest point corresponding to the fused interest point is further obtained, wherein the supplementary interest point is also an original interest point, that is, the data corresponding to the fused interest point is further collected to obtain a new original interest point as the supplementary interest point, and the fused interest point is fused again by using the supplementary interest point to update the information of the fused interest point, so that the information of the fused interest point is more accurate.

[0045] The method of the embodiment of the present disclosure fuses original interest points of different sources to obtain fused interest points, so as to improve the accuracy of the fused interest points. After obtaining the fused interest points, it is further judged whether the fused interest points can be online. For the fused interest points that do not meet the online condition, data is further collected to obtain corresponding supplementary original interest points, and the supplementary original interest points and the fused interest points are further fused, so as to improve the accuracy of the fused interest points.

[0046] On the basis of the above embodiment, first, how to obtain the fused interest points is described.

[0047] A plurality of original data is obtained, and the plurality of original data is obtained through a plurality of data collection methods. The original data is preprocessed according to a preset data format to obtain information of a plurality of original interest points.

[0048] For example, the original data is WIFI information, logistics information, Internet information, street view pictures and the like. The original data is preprocessed to extract information of the original interest points. The information of the original interest points can be in a preset data format, for example, the information includes a name, an address, a coordinate, a city and the like. For example, the address in the logistics information is geocoded to determine the coordinate of the original interest point.

[0049] For example, the original data includes a text containing an address. The information of the interest point is extracted from the text. For example, the CRF model can be used. The text is input into the model to obtain the corresponding recognition result. In the embodiment of the present disclosure, on the basis of the CRF, the Ernie pre-training model is introduced to add the natural language processing knowledge obtained based on unsupervised training to the input layer, so that the words / characters with strong relevance in the natural language understanding field are introduced into the input layer coding. The best expression coding of the input text is obtained as the output of the Ernie pre-training model. The output of the Ernie pre-training model is used as the input of the subsequent CRF model, so as to improve the accuracy of extracting the information of the interest point from the address.

[0050] After obtaining the information of the plurality of original interest points, it is determined whether there is a first fused interest point matched with the original interest point in the fusion information library according to the information of the original interest point. If there is, the first fused interest point is updated according to the information of the original interest point to obtain a fused interest point. If there is not, the original interest point is determined as a fused interest point in the fusion information library.

[0051] In the embodiments of the present disclosure, the information of the original interest points can be stored in an original interest point library, the information of the fused interest points can be stored in a fused interest point library, and the association relationship between the original interest points and the fused interest points, i.e., the corresponding relationship between one fused interest point and one or more original interest points, can be stored in a relationship library.

[0052] After each original interest point is obtained, it is determined according to the information of the original interest point whether there is a matched first fused interest point in the fused interest point library, i.e., whether the first fused interest point and the original interest point correspond to the same interest point. If there is, the first fused interest point can be further fused by using the information of the original interest point to obtain a fused interest point. If there is not, it indicates that the original interest point can be a new interest point, and the original interest point can be taken as a fused interest point.

[0053] For example, the first fused interest point is obtained by fusing the original interest point 1 and the original interest point 2. After the original interest point 3 is obtained in this step, it is determined that the original interest point 3 and the first fused interest point match. Then, the first fused interest point is updated by using the original interest point 3, i.e., the original interest point 1, the original interest point 2 and the original interest point 3 are fused together to obtain a fused interest point.

[0054] How to determine whether there is a matched first fused interest point in the fused interest point library according to the information of the original interest point is described.

[0055] The map is divided into a plurality of grids, and the surrounding fused interest points in the surrounding grids of the original interest point are obtained. The similarity of each item of information in the information of the original interest point and the information of each to-be-compared fused interest point is determined, the similarity score of each to-be-compared fused interest point is determined, and the to-be-compared fused interest point with the highest similarity score is determined. If the similarity score of the to-be-compared fused interest point with the highest similarity score is greater than a second preset value, the to-be-compared fused interest point with the highest similarity score is determined as the first fused interest point.

[0056] In the above method, the first step is the recall of the to-be-compared fused interest point, and the second step is the similarity comparison between each to-be-compared fused interest point and the original interest point, the to-be-compared fused interest point with the highest similarity score is selected as the most similar fused interest point, and whether the two interest points are the same interest point is determined according to the similarity score, i.e., whether the most similar fused interest point is the first fused interest point matched with the original interest point.

[0057] In the first step, the grid recall method can be used, the map is divided into N*N grids, and the fused interest points in the surrounding grids of the original interest point are recalled as the to-be-compared fused interest point set.

[0058] In the second step, the name similarity, address similarity, phone similarity, coordinate distance, and category consistency between the original interest point and each to-be-compared fused interest point are calculated, and these features are input into the GBDT model to obtain the similarity score between the original interest point and each to-be-compared fused interest point.

[0059] Optionally, for the original interest point or the to-be-compared fused interest point with the data source being WIFI information, new features can be added as the input of the GBDT model due to the particularity of the WIFI information (for example, the name is in pinyin, the first letter of pinyin, special suffix, and prefix). Optionally, the features include whether the name of the to-be-compared fused interest point contains Chinese, whether the name of the to-be-compared fused interest point contains letters or numbers, whether the name of the original interest point contains Chinese, whether the name of the original interest point contains letters or numbers, the effective length of the name of the to-be-compared fused interest point, the effective length of the name of the original interest point, the number of matched bigrams, the number of matched trigrams, whether the first keyword is matched, whether the tail keyword is matched, the number of matched characters, the length of the longest common sequence (LCS), and the Euclidean distance. These features and the features in the previous paragraph are input into the GBDT model to obtain the similarity score between the original interest point and each to-be-compared fused interest point. Thus, the matching accuracy of the interest point is improved.

[0060] The following further describes how to fuse the original interest point and the first fused interest point after the first fused interest point that matches the original interest point is determined.

[0061] As described above, the information of the interest point can include the name, address, or coordinate, and different fusion methods can be used for different information.

[0062] Optionally, the information of the original interest point includes the name or address, the information of the original interest point is parsed to obtain at least one parsed sub-information, the score of the information of the original interest point is determined according to the weight of the sub-information, the information with the highest score in the information of the original interest point and the information of at least one first original interest point corresponding to the first fused interest point is determined as the information of the fused interest point, and the fused interest point is obtained.

[0063] The sub-information obtained by parsing the name of the original interest point includes, for example, a main component, a child component, a branch component, a symbol, a core word, a suffix, a business scope, a geographical location, etc. For example, the name of YY glasses (XX road store) is parsed to obtain the main component YY glasses, the core word of the main component YY, the business scope glasses, the branch component XX road store, the core word of the branch component XX road, and the suffix store. For each sub-information, a weight can be preset, so that after parsing, the score of the name of the original interest point can be obtained by weighting the sub-information included in the parsing result and the corresponding weight. Still taking the first fused interest point obtained by fusing the original interest point 1 and the original interest point 2 as an example, the original interest point 1 and the original interest point 2 are the first original interest point. After obtaining the original interest point 3, the score of the name of the original interest point 3 is determined, compared with the score of the name of the original interest point 1 and the score of the name of the original interest point 2, and the name with the highest score is determined as the name of the fused interest point, thereby improving the accuracy of the name. The score of the name of the original interest point 1 and the score of the name of the original interest point 2 can be obtained when the two original interest points are fused to obtain the first fused interest point.

[0064] The sub-information obtained by parsing the address of the original interest point includes, for example, a province, a city, a district, a county, a town, a township, a street, a road, an alley, an administrative village, a community, a natural village, a house number, a village group, a building number, an office building (unit), a unit number, and a house number. For each sub-information, a weight can be preset, so that after parsing, the score of the address of the original interest point can be obtained by weighting the sub-information included in the parsing result and the corresponding weight. Still taking the first fused interest point obtained by fusing the original interest point 1 and the original interest point 2 as an example, the original interest point 1 and the original interest point 2 are the first original interest point. After obtaining the original interest point 3, the score of the address of the original interest point 3 is determined, compared with the score of the address of the original interest point 1 and the score of the address of the original interest point 2, and the address with the highest score is determined as the address of the fused interest point, thereby improving the accuracy of the address. The score of the address of the original interest point 1 and the score of the address of the original interest point 2 can be obtained when the two original interest points are fused to obtain the first fused interest point.

[0065] Optionally, the information of the original interest point includes coordinates, the average value of the coordinates of the second original interest point with the same coordinate source in the first original interest point corresponding to the original interest point and the first fused interest point is determined, and the coordinates of the second original interest point closest to the average value of the coordinates are determined as the coordinate source coordinates. The different coordinate sources are clustered, and the coordinates closest to the cluster center in the cluster with the most elements obtained after clustering are determined as the coordinates of the fused interest point, so as to obtain the fused interest point.

[0066] For example, the first original interest point corresponding to the first fused interest point has 10, after obtaining the original interest point of this step, among the 10 first original interest points and the original interest point, assuming that the coordinates of 3 second original interest points are all from WIFI information, then the coordinates of the 3 second original interest points are weighted and averaged to obtain the coordinate average value, and the coordinates of the second original interest point closest to the coordinate average value among the 3 second original interest points are determined as the coordinates of the coordinate source. For other coordinate sources, the type method is adopted, for example, for the second original interest point whose coordinate source is logistics information, the coordinates of the coordinate source are also determined, and the coordinates of different coordinate sources are clustered by DBSCAN. The cluster with the most elements obtained after clustering may have higher coordinate accuracy, so the coordinates closest to the cluster center in the cluster are determined as the coordinates of the fused interest point. By merging the coordinates of the original interest points with the same coordinate source, the problem of coordinate distortion caused by data enrichment from the same source can be avoided, and by selecting the coordinates closest to the mean / center point instead of the mean / center point itself, the situation that the calculation result falls on the road / water surface can be avoided.

[0067] By matching and fusing the original interest points, the fused interest points are obtained, and the accuracy of the interest points is improved. In addition, after obtaining the fused interest points, it is further needed to judge whether they can be online.

[0068] In the embodiments of the present disclosure, the GBDT model is used to judge whether the fused interest point can be online, and the time information corresponding to each information source of the information of the fused interest point is determined to determine whether the fused interest point satisfies the online condition. The input features of the model include a plurality of binary basic features, and each basic feature represents whether the original interest point of the corresponding data source is associated with the fused interest point. In addition, for the original interest point with strong time correlation of data source, such as the original interest point with data source of logistics information or WIFI information, the input features further include state features, which are used to represent whether there is an original interest point with data source of logistics information or WIFI information within a certain time. For example, if there is logistics information in the previous 1 month and the previous 5 months, the corresponding state feature is [1, 0, 0, 0, 1, 0, 0, 0, 0, 0]. Through this method, when judging online, the state change caused by the time of the data source is considered, and the time information of the data source is transmitted into the model as a feature, which can make the model learn the influence of the time information on the accuracy of the fused interest point, thereby ensuring the accuracy of the online judgment.

[0069] After the online judgment, for the fused interest point that does not satisfy the online condition, the update factor thereof, that is, the value or importance of the fused interest point, needs to be determined.

[0070] In the embodiments of the present disclosure, the values of multiple dimensions of the fused interest point are calculated, and an updating factor of the fused interest point is obtained through weighted summation. The multiple dimensions are described below. For example, the updating factor of the fused interest point is determined according to at least one of information of the fused interest point, a matching result of the fused interest point and a preset interest point library, a correlation degree of the fused interest point and an online interest point, and a click volume (PV) of a surrounding interest point of the fused interest point. The calculation through multiple dimensions improves the accuracy of the updating factor, that is, the accuracy of the value of the fused interest point.

[0071] Basic verification grading: The goal of basic verification grading is to calculate the value of the fused interest point based on information (name, address, coordinates, etc.) of the fused interest point itself. Through judging multiple dimensional features such as whether the information of the fused interest point contains random codes, whether it contains special characters, whether it contains sensitive words, whether it is empty, whether there is a mismatch judgment of parentheses, whether the coordinates match the city, and whether the coordinates exist in a road crushing and falling water case, a value score is calculated through a GBDT model.

[0072] Positive and negative example library construction: The positive example library is a high-accuracy interest point set mined from the full amount of interest points on the map line; the negative example library is a set of interest points that are historically offline and manually verified offline; the fused interest point is matched with the positive and negative example libraries, and according to the set to which the interest point with the highest matching score belongs, the positive example library or the negative example library, if the set to which the interest point with the highest matching score belongs is the positive example library, the value score is higher. The positive and negative example libraries are preset and can be dynamically updated.

[0073] Online search analysis: From the dimension of online search, the correlation degree between the fused interest point and the online interest point in the map is determined, and through outlier analysis, it is determined whether the fused interest point is an outlier. If the fused interest point is not an outlier, the PV value of the fused interest point can also be determined according to the PV of the interest point associated with the fused interest point on the line.

[0074] Nearby POI analysis: The PV value of the fused interest point is estimated from the spatial location dimension. Assuming that the spatial location of the fused POI is (x, y), the PV value of the fused interest point is determined by calculating the PV average of the interest points within a certain distance around the fused interest point, for example, the interest points within the range of [x-200, x+200][y-200, y+200].

[0075] Cutting block posterior analysis: The information of the full amount of interest points on the line, such as source, city, industry, and multiple dimensions, is analyzed in blocks to determine the influence of the source, city, and industry on the PV of the interest point, thereby determining the PV value of the fused interest point.

[0076] Through the above method, the value of the fused interest point in multiple dimensions is determined, and a weighted sum is performed to obtain an update factor of the fused interest point. The greater the update factor, the greater the value or importance of the fused interest point.

[0077] For the fused interest point with an update factor greater than the first preset value, a data collection method is determined according to the information of the fused interest point, and a supplementary interest point corresponding to the fused interest point obtained through the data collection method is acquired. For the fused interest point with an update factor not greater than the first preset value, the fused interest point can be discarded.

[0078] For example, the data collection method includes Internet crawling, regional collection, manual operation, etc. In practice, it can be determined according to the information of the fused interest point which data collection method to use to ensure collection efficiency, collection effect, and reduce collection cost.

[0079] For example, based on the name and geographic location, more information related to the fused interest point is obtained from the Internet to obtain a supplementary interest point, so that the supplementary interest point and the fused interest point can be further fused to meet the online condition.

[0080] For example, if there are a large number of fused interest points that do not meet the online condition in a street, a region, or even a city, regional collection can be performed through dispatching crowdsourcing and collection vehicles to obtain a large number of supplementary interest points.

[0081] For example, for the fused interest point that does not meet the online condition, a large number of users have searched in the map and have not obtained the desired result, the fused interest point can be distributed to manual operation. The scene is described.

[0082] In the search log, a series of user search behaviors and corresponding time points are recorded, such as 'input search query in search box', 'click interest point in search return list page', 'click to navigate to this', 'click to view interest point detail page', and basic information such as the city where the user initiates the search behavior and the city in the view on the map. Through analysis of the search log, the corresponding interest point may be missing on the map, which leads to poor user search satisfaction, and the keywords are mined in the search query input by the user.

[0083] By analyzing the sequence of relevant behaviors of a user in a complete search process, it is determined whether the user is satisfied with this search process. For example, in a search process of 'inputting a search query in a search box', 'clicking an interest point in a search return list page', and 'clicking to select navigation to the interest point', the user's search satisfaction is probably high, because the user may have found the desired search query result and selected navigation to the interest point. On the contrary, in a search process of 'inputting a search query in a search box', 'clicking to flip a page in a search list page', and 'quitting the app', the user's search satisfaction is probably low, because the user may not have found the desired interest point in the first page of search results, and after flipping the page, the user still did not find the desired interest point, and the user did not perform a behavior of 'clicking to select navigation to the interest point' to indicate satisfaction with the search result, but selected 'quitting the app'.

[0084] The triple file obtained in the previous step is further simplified. A user may input 'how should I go to work as soon as possible', which is not an interest point. Such a query is treated as a query with low search satisfaction in the user behavior analysis process, but actually such a query does not help to supplement the missing interest points of the map. In the embodiment of the present disclosure, a deep learning method is used to simplify the triple, that is, to denoise.

[0085] The task of the deep learning model is a text classification, that is, a binary classification of whether the query is an interest point. However, unlike the intuitive interest point binary classification, the task needs to find a query with low search satisfaction but potential to be solved. If a user searches for 'Southwest United University', which is indeed an interest point but only exists in history and does not exist in the real world, it does not meet the requirements.

[0086] To this end, in the embodiment of the present disclosure, when training the deep learning model, when labeling the training sample, the triple file generated by analyzing the user behavior is analyzed, and the'retrieval satisfaction rate' is represented by subtracting the 'number of times of query retrieval with poor satisfaction' from the 'total number of query retrieval on the same day' and dividing by the 'total number of query retrieval on the same day'. If the query satisfies 'the retrieval satisfaction rate' starts very low, but suddenly rises over time or 'the retrieval unsatisfactory rate' is always high, then the sample is labeled as a positive sample; on the contrary, if the'retrieval satisfaction rate' of the query is always low, then the sample is labeled as a negative sample, for example, the example in Table 1.

[0087] Table 1

[0088]

[0089] The above-mentioned way is used to label the training sample, and the deep learning model is trained, so that the deep learning model can accurately classify whether the query is a point of interest, thereby simplifying the triple and obtaining the target query. The points of interest in the target query are matched with the fusion points of interest that do not meet the online condition, and the matched fusion points of interest can be distributed to manual operation to collect data to obtain supplementary points of interest, and the fusion points of interest are updated to meet the online condition.

[0090] In the method of the embodiment of the present disclosure, the historical points of interest obtained in the past are stored in the corresponding original point of interest library or fusion point of interest library, and the historical points of interest can be verified by the newly obtained original points of interest, thereby improving the utilization rate of historical data and avoiding data loss. In addition, periodic deletion can also be used to avoid the occupation of storage space by too much invalid data.

[0091] If there is a second fusion point of interest in the fusion information library that has not been matched successfully with any original point of interest within a preset time period, the second fusion point of interest is deleted.

[0092] The fusion information library stores a large number of fusion points of interest obtained by fusing one original point of interest or multiple original points of interest. If the second fusion point of interest has no corresponding original point of interest matched or fused with it for a long time, the second fusion point of interest may be incorrect or non-existent. Therefore, the second fusion point of interest can be deleted to avoid storing incorrect data for a long time and occupying storage space.

[0093] Figure 3 It is a structural schematic diagram of a point of interest processing device provided according to an embodiment of the present disclosure. As shown in Figure 3 The point of interest processing device 300 includes:

[0094] The acquisition module 301 is configured to acquire a fused interest point, the fused interest point being obtained by data fusion on information of at least one original interest point.

[0095] The judgment module 302 is configured to judge whether the fused interest point meets an online condition.

[0096] The update module 303 is configured to, if the fused interest point does not meet the online condition, determine an update factor of the fused interest point, and acquire a supplementary interest point corresponding to the fused interest point according to the information of the fused interest point, so as to update the fused interest point according to the supplementary interest point, if the update factor is greater than a first preset value.

[0097] In an embodiment, the update module 303 comprises:

[0098] The first determination unit is configured to determine a data collection manner according to the information of the fused interest point.

[0099] The first acquisition unit is configured to acquire a supplementary interest point corresponding to the fused interest point obtained by the data collection manner.

[0100] In an embodiment, the judgment module 302 comprises:

[0101] The second determination unit is configured to determine that the fused interest point meets the online condition according to a source of each item of information in the information of the fused interest point and time information corresponding to the source of each item of information.

[0102] In an embodiment, the acquisition module 301 comprises:

[0103] The second acquisition unit is configured to acquire information of a plurality of original interest points.

[0104] The matching unit is configured to determine, according to the information of the original interest point, whether there is a first fused interest point matching the original interest point in a fused information library.

[0105] The fusion unit is configured to, if there is the first fused interest point, update the first fused interest point according to the information of the original interest point to obtain the fused interest point, and if there is not the first fused interest point, determine the original interest point as the fused interest point in the fused information library.

[0106] In an embodiment, the matching unit comprises:

[0107] The first acquisition subunit is configured to divide a map into a plurality of grids, and acquire a to-be-compared fused interest point in a grid around the original interest point.

[0108] The matching subunit is configured to determine the similarity between the information of the original interest point and each item of information of each to-be-compared fused interest point, determine a similarity score of each to-be-compared fused interest point, and determine a to-be-compared fused interest point with the highest similarity score, and if the similarity score of the to-be-compared fused interest point with the highest similarity score is greater than a second preset value, determine the to-be-compared fused interest point with the highest similarity score as the first fused interest point.

[0109] In an embodiment, the information includes a name or an address.

[0110] The fusion unit includes:

[0111] The parsing subunit is configured to parse the information of the original interest point to obtain at least one parsed sub-information.

[0112] The fusion subunit is configured to determine a score of the information of the original interest point according to the weight of the sub-information, and determine the information of the fused interest point as the information of the fused interest point by using the information of the original interest point and the information of at least one first original interest point corresponding to the first fused interest point with the highest score.

[0113] In an embodiment, the information includes a coordinate.

[0114] The fusion unit includes:

[0115] The first determining subunit is configured to determine a coordinate average value of the second original interest points with the same coordinate source among the original interest point and the first original interest points corresponding to the first fused interest point, and determine the coordinate of the second original interest point closest to the coordinate average value as the coordinate of the coordinate source.

[0116] The second determining subunit is configured to cluster the coordinates of different coordinate sources, and determine the coordinate closest to the cluster center in the cluster with the most elements after clustering as the coordinate of the fused interest point to obtain the fused interest point.

[0117] In an embodiment, the update module 303 includes:

[0118] The third determining unit is configured to determine an update factor of the fused interest point according to at least one of the information of the fused interest point, the matching result of the fused interest point and the preset interest point library, the correlation degree of the fused interest point and the online interest point, and the click volume of the interest points around the fused interest point.

[0119] In an embodiment, the second obtaining unit includes:

[0120] The second obtaining subunit is configured to obtain a plurality of original data, and the plurality of original data is obtained through a plurality of data collection methods.

[0121] The preprocessing subunit is configured to preprocess the original data according to a preset data format to obtain information of a plurality of original interest points.

[0122] In an embodiment, the interest point processing apparatus 300 further comprises:

[0123] The deleting module is configured to delete the second fused interest point if the second fused interest point that has not been matched with any original interest point successfully within a preset time period exists in the fused information library.

[0124] The apparatus of the embodiments of the present disclosure can be used to execute the interest point processing method in the method embodiments described above, and the implementation principles and technical effects are similar, which will not be described here again.

[0125] According to the embodiments of the present disclosure, the present disclosure further provides an electronic device and a non-transitory computer-readable storage medium storing computer instructions.

[0126] According to the embodiments of the present disclosure, the present disclosure further provides a computer program product, which comprises a computer program stored in a readable storage medium, at least one processor of an electronic device can read the computer program from the readable storage medium, and the at least one processor executes the computer program to make the electronic device execute the scheme provided in any of the embodiments described above.

[0127] Figure 4 is a schematic block diagram of an electronic device for implementing the interest point processing method of the embodiments of the present disclosure. The electronic device is intended to represent various forms of digital computers, such as laptops, desktops, tablets, personal digital assistants, servers, blade servers, mainframes, and other appropriate computers. The electronic device can also represent various forms of mobile devices, such as personal digital processors, cellular telephones, smart phones, wearable devices, and other similar computing devices. The components shown here, their connections and relationships, and their functions, are merely examples, and are not intended to limit the implementations described and / or claimed in this document.

[0128] As shown in Figure 4 The electronic device 400 includes a computing unit 401 that can perform various appropriate actions and processes according to computer programs stored in a read-only memory (ROM) 402 or loaded into a random access memory (RAM) 403 from a storage unit 408. Various programs and data required for the operation of the device 400 can also be stored in the RAM 403. The computing unit 401, the ROM 402, and the RAM 403 are connected to each other through a bus 404. An input / output (I / O) interface 405 is also connected to the bus 404.

[0129] A plurality of components in the device 400 are connected to the I / O interface 405, including: an input unit 406, such as a keyboard, a mouse, etc.; an output unit 407, such as various types of displays, speakers, etc.; a storage unit 408, such as a magnetic disk, an optical disk, etc.; and a communication unit 409, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 409 allows the device 400 to exchange information / data with other devices through a computer network, such as the Internet, and / or various telecommunication networks.

[0130] The computing unit 401 can be various general and / or special-purpose processing components with processing and computing capabilities. Some examples of the computing unit 401 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various special-purpose artificial intelligence (AI) computing chips, various computing units running machine learning model algorithms, a digital signal processor (DSP), and any appropriate processor, controller, microcontroller, etc. The computing unit 401 performs various methods and processes described above, such as the processing method of points of interest. For example, in some embodiments, the processing method of points of interest can be implemented as a computer software program, which is tangibly embodied in a machine-readable medium, such as the storage unit 408. In some embodiments, part or all of the computer program can be loaded and / or installed onto the device 400 via the ROM 402 and / or the communication unit 409. When the computer program is loaded onto the RAM 403 and executed by the computing unit 401, one or more steps of the processing method of points of interest described above can be performed. Alternatively, in other embodiments, the computing unit 401 can be configured to perform the processing method of points of interest by any other appropriate means, such as by means of firmware.

[0131] The various implementations of the systems and techniques described above herein can be realized in a digital electronic circuit system, an integrated circuit system, a field programmable gate array (FPGA), an application specific integrated circuit (ASIC), an application specific standard product (ASSP), a system on a chip system (SOC), a complex programmable logic device (CPLD), computer hardware, firmware, software, and / or combinations thereof. These various implementations can include implementation in one or more computer programs that are executable and / or interpretable on a programmable system including at least one programmable processor, which can be special or general purpose, coupled to receive data and instructions from, and to transmit data and instructions to, a storage system, at least one input device, and at least one output device.

[0132] Program code for carrying out methods of the present disclosure can be written in any combination of one or more programming languages. The program code can be provided to a processor or controller of a general purpose computer, special purpose computer, or other programmable data processing apparatus to produce a machine, such that the program code, when executed by the processor or controller, produces the functions / operations specified in the flowcharts and / or block diagrams. The program code can be executed entirely on a machine, partially on a machine, partially on a machine as a standalone software package, or entirely on a remote machine or server.

[0133] In the context of the present disclosure, a machine-readable medium can be a tangible medium that contains or stores a program for use by or in connection with an instruction execution system, apparatus, or device. The machine-readable medium can be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium can include but is not limited to an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any suitable combination of the foregoing. More specific examples of the machine-readable storage medium will include one or more lines of electrical connections, portable computer disks, hard disk drives, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or Flash memory), optical fibers, portable compact disc read-only memories (CD-ROMs), optical storage devices, magnetic storage devices, or any suitable combination of the foregoing.

[0134] To provide for interaction with a user, the systems and techniques described here can be implemented on a computer having a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user and a keyboard and a pointing device (e.g., a mouse or a trackball) by which the user can provide input to the computer. Other kinds of devices can be used to provide for interaction with a user as well; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form, including acoustic, speech, or tactile input.

[0135] The systems and techniques described here can be implemented in a computing system that includes a back end component (e.g., as a data server), or that includes a middleware component (e.g., an application server), or that includes a front end component (e.g., a user computer having a graphical user interface or a Web browser through which a user can interact with an implementation of the systems and techniques described here), or any combination of such back end, middleware, or front end components. The components of the system can be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include a local area network (LAN), a wide area network (WAN), and the Internet.

[0136] The computer system can include clients and servers. This relationship can be. The servers are typically remote from the clients with the interactions typically taking place over a communication network. The relationship between a client and a server is one of client-server relationship. The server can be a cloud server, also known as cloud computing server or cloud host, which is a host product in the cloud computing service system. The server can also be a server of a distributed system or a server combined with a blockchain.

[0137] It should be understood that the various forms of flow shown above can be re-ordered, added to, or deleted from without departing from the scope of the present disclosure. For example, the steps recited in the present application can be executed in parallel, in series, or in a different order, without departing from the desired results of the technical solutions disclosed in the present disclosure, and the present disclosure is not limited herein.

[0138] The above detailed description does not constitute a limitation on the protection scope of the present disclosure. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent replacements, and improvements within the spirit and principles of the present disclosure should be included in the protection scope of the present disclosure.

Claims

1. A method for processing points of interest, comprising: Acquire fused points of interest, wherein the fused points of interest are obtained by data fusion of information from at least one original point of interest; Determine whether the fused points of interest meet the online conditions; If the conditions are not met, the update factor of the fused interest point is determined, and when the update factor is greater than the first preset value, the supplementary interest point corresponding to the fused interest point is obtained according to the information of the fused interest point, so as to update the fused interest point according to the supplementary interest point. Different fusion methods are used to obtain the fused interest points from information on different points of interest. When the information includes coordinates, obtaining the fused points of interest includes: Based on the information of the original points of interest, when it is determined that there is a first fused point of interest in the fused information database that matches the original points of interest, the average coordinates of the second original points of interest with the same coordinate source as the original points of interest and the first fused points of interest are determined, and the coordinates of the second original points of interest that are closest to the average coordinates are determined as the coordinates of the coordinate source. Clustering coordinates from different coordinate sources, the coordinates closest to the cluster center in the cluster with the most elements are determined as the coordinates of the fused interest point, so as to obtain the fused interest point; When the information includes a name or address, obtaining the fused points of interest includes: The information of the original points of interest is parsed to obtain at least one sub-information after parsing; According to the weight of the sub-information, the score of the original interest point information is determined, and the information with the highest score among the original interest point information and the information of at least one first original interest point corresponding to the first fused interest point is determined as the information of the fused interest point, so as to obtain the fused interest point.

2. The method according to claim 1, wherein, The step of obtaining supplementary points of interest corresponding to the fused points of interest based on the information of the fused points of interest includes: Based on the information of the fused points of interest, the data acquisition method is determined; Obtain supplementary points of interest corresponding to the fused points of interest obtained through the data acquisition method.

3. The method according to claim 1 or 2, wherein, The determination of whether the fused points of interest meet the online conditions includes: Based on the source of each piece of information in the information of the fused points of interest, and the time information corresponding to the source of each piece of information, it is determined that the fused points of interest meet the online conditions.

4. The method according to claim 3, wherein, The method further includes: determining, based on the information of the original points of interest, whether there exists a first fused point of interest that matches the original points of interest in the fused information database; If it does not exist, the original point of interest will be identified as the fused point of interest in the fused information database.

5. The method according to claim 4, wherein, The step of determining whether a first fused interest point matching the original interest point exists in the fusion information database based on the information of the original interest point includes: The map is divided into multiple grids, and the points of interest to be compared and fused are obtained from the grids surrounding the original points of interest; The similarity between the information of the original interest point and the information of each interest point to be compared and fused is determined. The similarity score of each interest point to be compared and fused is determined. The interest point to be compared and fused with the highest similarity score is determined. If the similarity score of the interest point to be compared and fused with the highest similarity score is greater than a second preset value, then the interest point to be compared and fused with the highest similarity score is determined as the first fused interest point.

6. The method according to any one of claims 1-2 and 4-5, wherein, The step of determining the update factor for the fused interest points includes: The update factor of the fused interest point is determined based on at least one of the following: the information of the fused interest point, the matching result between the fused interest point and the preset interest point library, the correlation between the fused interest point and the online interest points, and the click volume of the interest points surrounding the fused interest point.

7. The method according to claim 4 or 5, wherein, The acquisition of information from multiple original points of interest includes: Multiple raw data points are acquired, which are obtained through various data acquisition methods. The original data is preprocessed according to a preset data format to obtain information on the multiple original points of interest.

8. The method according to claim 4 or 5, further comprising: If a second fused point of interest exists in the fused information database that has not been successfully matched with any original point of interest within a preset time period, then the second fused point of interest is deleted.

9. An apparatus for processing points of interest, comprising: The acquisition module is used to acquire fused points of interest, wherein the fused points of interest are obtained by data fusion of information from at least one original point of interest; The judgment module is used to determine whether the fused points of interest meet the online conditions; An update module is used to determine the update factor of the fused interest point if the condition is not met, and when the update factor is greater than a first preset value, obtain the supplementary interest point corresponding to the fused interest point based on the information of the fused interest point, so as to update the fused interest point based on the supplementary interest point. In this process, different fusion methods are used to obtain the fused interest points from information on different points of interest. When the information includes coordinates, the acquisition module includes a fusion unit. The fusion unit includes: The first determining subunit is configured to, based on the information of the original point of interest, when determining that there is a first fused point of interest in the fused information database that matches the original point of interest, determine the average coordinate of the second original points of interest whose coordinate sources are consistent with those of the original point of interest and the first fused point of interest, and determine the coordinate of the second original point of interest that is closest to the average coordinate as the coordinate source coordinate; The second determining subunit is used to cluster coordinates from different coordinate sources, and to determine the coordinates of the fused interest point as the coordinates of the cluster with the most elements obtained after clustering, which are closest to the cluster center. When the information includes a name or address, the fusion unit includes: The parsing subunit is used to parse the information of the original interest point to obtain at least one sub-information after parsing; The fusion subunit is used to determine the score of the information of the original interest point according to the weight of the sub-information, and to determine the information with the highest score among the information of the original interest point and the information of at least one first original interest point corresponding to the first fused interest point as the information of the fused interest point, so as to obtain the fused interest point.

10. The apparatus according to claim 9, wherein, The update module includes: The first determining unit is used to determine the data acquisition method based on the information of the fused points of interest; The first acquisition unit is used to acquire supplementary points of interest corresponding to the fused points of interest obtained through the data acquisition method.

11. The apparatus according to claim 9 or 10, wherein, The judgment module includes: The second determining unit is used to determine whether the fused interest point meets the online conditions based on the source of each piece of information in the information of the fused interest point and the time information corresponding to the source of each piece of information.

12. The apparatus according to claim 11, wherein, The acquisition module includes: The matching unit is used to determine, based on the information of the original point of interest, whether there exists a first fused point of interest that matches the original point of interest in the fusion information database; A fusion unit is used to determine the original point of interest as the fused point of interest in the fusion information database if the original point of interest does not exist.

13. The apparatus according to claim 12, wherein, The matching unit includes: The first acquisition subunit is used to divide the map into multiple grids and acquire the points of interest to be compared and fused in the grids surrounding the original points of interest; The matching subunit is used to determine the similarity between the information of the original interest point and the information of each interest point to be compared and fused, determine the similarity score of each interest point to be compared and fused, and determine the interest point to be compared and fused with the highest similarity score. If the similarity score of the interest point to be compared and fused with the highest similarity score is greater than a second preset value, then the interest point to be compared and fused with the highest similarity score is determined as the first fused interest point.

14. The apparatus according to any one of claims 9-10 and 12-13, wherein, The update module includes: The third determining unit is used to determine the update factor of the fused interest point based on at least one of the following: the information of the fused interest point, the matching result between the fused interest point and the preset interest point library, the correlation between the fused interest point and the online interest points, and the click volume of the interest points surrounding the fused interest point.

15. The apparatus according to claim 12 or 13, wherein, The acquisition module further includes a second acquisition unit, the second acquisition unit comprising: The second acquisition subunit is used to acquire multiple raw data, which are obtained through various data acquisition methods. The preprocessing subunit is used to preprocess the raw data according to a preset data format to obtain information about the multiple raw points of interest.

16. The apparatus according to claim 12 or 13, further comprising: The deletion module is used to delete the second fused interest point if there is a second fused interest point in the fused information database that has not been successfully matched with any original interest point within a preset time period.

17. An electronic device comprising: At least one processor; and memory that is communicatively connected to at least one processor; The memory stores instructions that can be executed by at least one processor, which, when executed by at least one processor, enables the at least one processor to perform the method of any one of claims 1-8.

18. A non-transitory computer-readable storage medium storing computer instructions for causing the computer to perform the method of any one of claims 1-8.

19. A computer program product comprising a computer program that, when executed by a processor, implements the method of any one of claims 1-8.

Images