A method and device for analyzing early warning events

Abstract

The present invention discloses a method and device for analyzing early warning events, and relates to the field of smart city technology. A specific implementation of the method includes: obtaining one or more early warning events; the early warning event indicates each occurrence location and early warning type; determining the early warning area corresponding to the occurrence location, and one or more interest point types included in the early warning area; for early warning events of the same early warning type: analyzing the type of interest point corresponding to the early warning event, and determining the target interest point type whose correlation with the early warning event is greater than a preset threshold; according to the target interest point type corresponding to each early warning event in the same early warning type, determining the interest point type combination corresponding to the early warning type. By associating the early warning type with the target interest point type combination, the early warning type and the target interest point type combination associated with it are transferable between different early warning events, thereby improving the analysis efficiency of early warning events.

Description

Technical Field

[0001] The present invention relates to the field of smart city technology, and in particular to an analysis method and device for early warning events. Background Art

[0002] In the field of urban governance, when a business item that needs to be governed occurs, urban governance personnel often start from the event that has occurred and analyze the causes or related factors behind it. However, the current analysis of related factors of warning events only analyzes the types of points of interest around a certain warning event and their relevance, for example, the relevance between the types of points of interest around a fire alarm point and the fire alarm event. This current analysis method is a single-point analysis and is not transferable between different warning events. Summary of the invention

[0003] In view of this, an embodiment of the present invention provides a method and device for analyzing a warning event, by analyzing the types of points of interest in the warning area corresponding to the warning event, obtaining a target point of interest type with a high correlation with the warning event, and then analyzing the combination of target point of interest types according to the warning type to obtain a target point of interest type combination corresponding to each warning type, thereby associating the warning type with the target point of interest type combination to obtain a point of interest type combination with a higher correlation with the warning type. Thus, the most relevant point of interest type combination corresponding to warning events belonging to the same warning type can be determined, so that according to the point of interest type combination of a certain region, it can be determined which types of warnings are required in the region, without the need to perform a single point analysis for each warning event, thereby making the warning type and the target point of interest type combination associated with it transferable between different warning events, thereby improving the analysis efficiency of the warning event.

[0004] To achieve the above objective, according to a first aspect of an embodiment of the present invention, a method for analyzing a warning event is provided.

[0005] The analysis method of warning events in an embodiment of the present invention includes: obtaining one or more warning events; the warning event indicates the occurrence location and the warning type; determining the warning area corresponding to each of the occurrence locations, and one or more interest point types included in the warning area; for the warning events of the same warning type: analyzing the interest point types corresponding to the warning events, and determining the target interest point types whose correlation with the warning events is greater than a preset threshold; determining the interest point type combination corresponding to the warning type according to the target interest point types corresponding to the respective warning events in the same warning type.

[0006] Optionally, the analysis of the interest point types corresponding to the warning events to determine the target interest point types having a correlation with the warning events greater than a preset threshold includes: determining the number of interest points of each interest point type in the warning area corresponding to each of the warning events; determining the distribution vector corresponding to each of the warning events based on the number of interest points and the total number of interest point types; and determining the target interest point type based on the distribution vector.

[0007] Optionally, determining the target interest point type according to the distribution vector includes: obtaining the distribution vector corresponding to each warning event in the same warning event type; forming a distribution matrix corresponding to the warning event type according to the distribution vector; based on the distribution matrix, determining the correlation between each interest point type and each warning event using the TF-IDF algorithm; and using the interest point type whose correlation with the warning event is greater than a preset threshold as the target interest point type of the warning event.

[0008] Optionally, based on the distribution matrix, the TF-IDF algorithm is used to determine the correlation between each interest point type and each of the warning events, including: determining the occurrence probability of each interest point type in the warning area corresponding to the warning event according to the distribution vector corresponding to the warning event; determining the inverse document probability of each interest point type in the warning event type corresponding to the warning event according to the distribution matrix; and determining the correlation between each interest point type and the warning event based on the occurrence probability and the inverse document probability.

[0009] Optionally, determining the combination of interest point types corresponding to the warning type includes: determining a target distribution matrix based on the distribution matrix and the target interest point types corresponding to each of the warning events; determining, based on the target distribution matrix and an Apriori algorithm, conditional probability values ​​of combinations of different interest point types appearing in the warning type; and determining, based on the conditional probability values, a combination of interest point types corresponding to the warning type from combinations of different interest point types.

[0010] Optionally, determining the target distribution matrix according to the distribution matrix and the target interest point type corresponding to each of the warning events includes: determining a Boolean matrix corresponding to the distribution matrix according to the target interest point type corresponding to each of the warning events; using the Boolean matrix as the target distribution matrix; wherein the distribution vector corresponding to each warning event in the Boolean matrix satisfies: the value of the target interest point type corresponding to the warning event is 1, and the values ​​of other interest point types are 0.

[0011] Optionally, determining the warning area corresponding to each occurrence location and one or more types of points of interest included in the warning area includes: determining the warning area corresponding to the warning event with the coordinates of the occurrence location of the warning event as the center and a preset distance as the radius; and obtaining the types of points of interest within the warning area from a database.

[0012] Optionally, the method also includes: receiving a warning query request, the warning query request indicating the location to be queried; determining one or more target points of interest corresponding to the location to be occurred; and determining and outputting the warning type of the warning event corresponding to the location to be occurred based on the point of interest types of the one or more target points of interest.

[0013] To achieve the above objective, according to a second aspect of an embodiment of the present invention, a device for processing a warning event is provided.

[0014] The processing device of the warning event of the embodiment of the present invention includes: an acquisition module, which is used to acquire one or more warning events; the warning event indicates the occurrence location and the warning type; a determination module, which is used to determine the warning area corresponding to each of the occurrence locations, and the type of one or more points of interest included in the warning area; an analysis module, which is used for: for the warning events of the same warning type: analyzing the type of point of interest corresponding to the warning event, and determining the target point of interest type whose correlation with the warning event is greater than a preset threshold; and determining the point of interest type combination corresponding to the warning type according to the target point of interest types corresponding to each warning event in the same warning type.

[0015] To achieve the above objective, according to a third aspect of an embodiment of the present invention, a warning event analysis device is provided.

[0016] The analysis device for warning events according to the embodiment of the present invention includes: one or more processors; a storage system for storing one or more programs; when the one or more programs are executed by the one or more processors, the one or more processors implement the analysis method for warning events according to the embodiment of the present invention.

[0017] To achieve the above objective, according to a fourth aspect of an embodiment of the present invention, a computer-readable medium is provided.

[0018] A computer program is stored on a computer-readable medium in the embodiment of the present invention. When the program is executed by a processor, the method for analyzing the warning event in the embodiment of the present invention is implemented.

[0019] An embodiment of the above invention has the following advantages or beneficial effects: In the embodiment of the present invention,

[0020] By analyzing the types of points of interest in the warning area corresponding to the warning event, the target points of interest types with a high correlation with the warning event are obtained, and then the combinations of target points of interest types are analyzed according to the warning type to obtain the target points of interest type combinations corresponding to each warning type, so as to associate the warning type with the target points of interest type combinations to obtain the points of interest type combinations with a higher correlation with the warning type. In this way, the most relevant points of interest type combinations corresponding to the warning events belonging to the same warning type can be determined, so that according to the points of interest type combinations in a certain area, it is possible to determine which types of warnings are required in the area, without the need to perform single-point analysis for each warning event, so that the warning type and the target points of interest type combinations associated with it are transferable between different warning events, thereby improving the analysis efficiency of the warning event.

[0021] The further effects of the above-mentioned non-conventional optional manner will be described below in conjunction with the specific implementation manner. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] The accompanying drawings are used to better understand the present invention and do not constitute an improper limitation of the present invention.

[0023] Figure 1 It is a schematic diagram of the main process of a method for analyzing a warning event according to an embodiment of the present invention;

[0024] Figure 2 It is a schematic diagram of the main process of determining the warning area corresponding to the occurrence location according to an embodiment of the present invention;

[0025] Figure 3 is a schematic diagram of the main process of determining the target interest point type corresponding to the warning event according to an embodiment of the present invention;

[0026] Figure 4 is a schematic diagram of a main process of further determining the type of a target point of interest according to an embodiment of the present invention;

[0027] Figure 5 is a schematic diagram of a main process of determining the relevance between each interest point type and each warning event according to an embodiment of the present invention;

[0028] Figure 6 is a schematic diagram of a main process of determining a combination of interest point types corresponding to an early warning type according to an embodiment of the present invention;

[0029] Figure 7 is a schematic diagram of the main process of determining the target distribution matrix according to an embodiment of the present invention;

[0030] Figure 8 It is a specific flow chart of a method for analyzing a warning event according to an embodiment of the present invention;

[0031] Fig. 9 This is a schematic diagram of a specific process of predicting a warning event that has not occurred according to an embodiment of the present invention;

[0032] Fig.10 It is a schematic diagram of main modules of a warning event analysis device according to an embodiment of the present invention;

[0033] Fig.11 An exemplary system architecture diagram in which embodiments of the present invention may be applied;

[0034] Fig.12 A schematic diagram of the structure of a computer system of a terminal device or a server suitable for implementing an embodiment of the present invention. DETAILED DESCRIPTION

[0035] The following is a description of exemplary embodiments of the present invention in conjunction with the accompanying drawings, including various details of the embodiments of the present invention to facilitate understanding, which should be considered as merely exemplary. Therefore, it should be recognized by those of ordinary skill in the art that various changes and modifications may be made to the embodiments described herein without departing from the scope and spirit of the present invention. Similarly, for clarity and conciseness, the description of well-known functions and structures is omitted in the following description.

[0036] According to a first aspect of an embodiment of the present invention, a method for analyzing a warning event applied to a server is provided.

[0037] In the existing warning analysis methods, the monitoring and warning system can provide users with the warning event surrounding analysis function, but it can only provide the most frequent point of interest (POI) type combination around the location of a certain warning event, and cannot be associated with the warning type of the warning event. In addition, in the case of different warning types, only single-point analysis can be performed based on a single warning type, which is not transferable and co-occurring, and has certain limitations. Therefore, this application realizes in-depth mining of the warning event surrounding analysis function by associating the point of interest type with the warning type.

[0038] Figure 1 FIG. 1 is a schematic diagram of the main process of a method for analyzing a warning event according to an embodiment of the present invention. Figure 1 As shown, the method mainly includes:

[0039] Step S101: Obtain one or more warning events; the warning event indicates the occurrence location and warning type;

[0040] Step S102: determining the warning area corresponding to each occurrence location and one or more types of points of interest included in the warning area;

[0041] For warning events of the same warning type, continue to execute step S103 and step S104;

[0042] Step S103: analyzing the types of interest points corresponding to the warning event, and determining the target interest point types whose correlation with the warning event is greater than a preset threshold;

[0043] Step S104: determining a combination of interest point types corresponding to the warning type according to the target interest point types corresponding to the warning events in the same warning type.

[0044] As a monitoring and management system for the city status, the monitoring and early warning system will receive and process the early warning data on the city governance business items generated during the city operation process, and send early warning messages to users. Among them, in the process of the monitoring and early warning system monitoring the city operation status, early warning events are usually generated. In an optional embodiment, the early warning event may include the location of occurrence (coordinate information), early warning type, early warning level, early warning time, and early warning source information.

[0045] Exemplarily, the location of the warning event can be specifically the longitude and latitude coordinates of the place where the warning event occurred. The warning type refers to the business areas involved in urban governance that require early prevention of risks, such as fire warnings, group rental warnings, traffic accident warnings, fatigue driving warnings, etc. The warning level can be divided into level one to level five from low to high. The higher the warning level, the more times accidents of this warning type occur and the greater the probability of their occurrence.

[0046] In an embodiment of the present invention, POI (Point of Interest) can be abstracted as a point geographical object, especially some geographical entities that are closely related to people's lives. For example, AA restaurant, BB school, CC hospital, DD supermarket, etc. Furthermore, each point of interest can contain various information, such as name, category, coordinates, classification, etc. The point of interest type is a classification of points of interest, which means a class of POIs with the same semantic meaning, regardless of the specific geographical location, such as schools, banks, restaurants, gas stations, hospitals, supermarkets, bus stops, etc. Among them, when school is used as the point of interest type, all schools in the warning area can be included, such as BB1 school, BB2 school, BB3 school, etc.

[0047] In an optional embodiment, if Figure 2 As shown, step S102 may further include:

[0048] Step S201: Taking the coordinates of the location where the warning event occurs as the center of the circle and a preset distance as the radius, determine the warning area corresponding to the warning event;

[0049] Step S202: Obtain the type of points of interest within the warning area from the database.

[0050] Exemplarily, the location where the warning event occurs is point a and the preset distance is 1 km, then the warning area is a circular area S(a) with point a as the center and a radius of 1 km, and the one or more point of interest types included in the warning area are the point of interest types corresponding to the points of interest included in the circular area S(a).

[0051] In an optional embodiment, the process of determining the target interest point type corresponding to the warning event in step S103 is as follows: Figure 3 As shown, further comprising:

[0052] Step S301: determining the number of points of interest of each point of interest type in the warning area corresponding to each warning event;

[0053] Step S302: determining a distribution vector corresponding to each warning event according to the number of points of interest and the total number of types of points of interest;

[0054] Step S303: Determine the type of the target interest point according to the distribution vector.

[0055] Specifically, the distribution vector of the type of points of interest in the corresponding warning area for each warning event can be expressed as Where n represents the total number of interest points types, and vec[i] represents the number of interest points of type i that appear in the warning area corresponding to the location of the current warning event. For example, the location of the warning event is point a, and the warning area is S(a). There are 6 types of interest points in the warning area S(a), and categories 1 to 6 are schools, banks, restaurants, gas stations, hospitals, and supermarkets, that is, n is 6. Among them, there are 1 school, 1 gas station, and 1 hospital, 3 banks, 5 restaurants, and 2 supermarkets. Then vec[1] represents the number of interest points belonging to the school, that is, vec[1] = 1, and vec[2] represents the number of interest points belonging to the bank, that is, vec[2] = 3. And so on. In the warning area corresponding to each warning event, each type of interest point has a corresponding number of interest points. Furthermore, the distribution vector corresponding to each warning event can be determined according to the number of points of interest and the total number of points of interest types, that is, the distribution vector corresponding to 1 school, 3 banks, 5 restaurants, 1 gas station, 0 hospitals, and 2 supermarkets is (1, 3, 5, 1, 0, 2).

[0056] For step S303, in an alternative embodiment, the TF-IDF (term frequency–inverse document frequency) algorithm can be used to analyze the types of the points of interest corresponding to the warning events. In the prior art, the TF-IDF algorithm is a commonly used weighting technique for information retrieval and data mining, and is often used to mine keywords in articles. TF represents term frequency (abbreviated as TF). A high term frequency in an article often refers to some common words without practical significance, such as "de" (的), "shi" (是), "le" (了), etc. IDF (Inverse Document Frequency, abbreviated as IDF) represents inverse document frequency. IDF gives smaller weights to common words, and its magnitude is inversely proportional to the commonness of a word. When TF (term frequency) and IDF (inverse document frequency) are multiplied, the TF-IDF value of a word can be obtained. By filtering the words in the document according to the TF-IDF value, the remaining words often have practical significance. After processing the words in the document with the TF-IDF algorithm, the probability of obtaining the keywords of the document will increase significantly. In the present invention, by using the TF-IDF algorithm to screen the types of the points of interest around the occurrence location of the warning event, some types of points of interest (noise) that often appear in the vast majority of geographical locations but have a small quantity can be excluded, and the types of the points of interest with a greater correlation with the warning type can be obtained, and further the types of the points of interest with a greater correlation with the warning event can be obtained.

[0057] In a further alternative embodiment, as Figure 4 shown, the process of determining the target type of the point of interest in step S303 specifically includes:

[0058] Step S401: Obtain the distribution vectors corresponding to each warning event in the same warning event type;

[0059] Step S402: According to the distribution vectors, form a distribution matrix corresponding to the warning event type;

[0060] Step S403: Based on the distribution matrix, use the TF-IDF algorithm to determine the correlation degree between each type of point of interest and each warning event;

[0061] Step S404: Use the type of the point of interest with a correlation degree greater than a preset threshold with the warning event as the target type of the point of interest for this warning event.

[0062] Among them, for step S402, illustratively, assuming that the warning event type is a fire type, then three warning events belonging to the fire type can be obtained, and the distribution vectors corresponding to each warning event can be obtained according to steps S301 to S303. For example, the distribution vector of warning event 1 is (1, 0, 3, 1, 0, 1), the distribution vector of warning event 2 is (2, 1, 4, 2, 0, 1), and the distribution vector of warning event 3 is (0, 0, 2, 0, 1, 1). The distribution vectors of the three warning events are combined to form a distribution matrix corresponding to the warning event type:

[0063] 1,0,3,1,0,1

[0064] 2,1,4,2,0,1

[0065] 0,0,2,0,1,1

[0066] Among them, the distribution matrix is ​​a mat matrix of numerical type, which can be specifically expressed as Where n represents the total number of interest point types, m represents the total number of multiple warning events belonging to the same warning type, and k represents the kth warning type. After obtaining the distribution matrix, the TF-IDF algorithm can be used to determine the correlation between each interest point type and each warning event. In the calculation process of the TF-IDF algorithm, two values ​​need to be determined, one is the frequency of the interest point (such as school) in the distribution vector of the warning event, and the other is the inverse document probability of the interest point (such as school) in the distribution matrix corresponding to the entire fire type.

[0067] Therefore, in an optional embodiment, the process of using the TF-IDF algorithm to determine the relevance between each interest point type and each of the warning events in step S403 is as follows: Figure 5 As shown, it may further include:

[0068] Step S501: determining the occurrence probability of each point of interest type in the warning area corresponding to the warning event according to the distribution vector corresponding to the warning event;

[0069] Step S502: determining the inverse document probability of each interest point type in the warning event type corresponding to the warning event according to the distribution matrix;

[0070] Step S503: Based on the occurrence probability and the inverse document probability, determine the correlation between each interest point type and the warning event.

[0071] Taking the above distribution matrix as an example, for warning event 1, when the interest point type is school, the probability of occurrence in warning event 1 is 1 / (1+0+3+1+0+1)=1 / 6, and the inverse document probability in the warning event type of fire type (i.e., the distribution matrix) is (log 1 / (2 / 3)+ε)). After obtaining the probability of occurrence of each interest point type in the warning area corresponding to the warning event and the inverse document probability in the warning event type corresponding to the warning event, the correlation can be solved according to the following formula (1):

[0072]

[0073] Among them, P j (i) represents the probability of occurrence of interest point type i in the warning area corresponding to the occurrence location j of the warning event; log(1 / P k (i)+ε) represents the inverse document probability of interest point type i in the warning area corresponding to the location of the kth warning event; where P k (i) represents the probability of the i-th type of interest point appearing in the warning area corresponding to the location where the k-th type of warning event occurs; ε represents an infinite decimal, which is set to a constant; I represents the set of interest point types, J represents the set of warning locations, and K represents the set of warning types.

[0074] After determining the correlation between each warning event and each POI type under the same warning type, in a further optional embodiment, the correlation is screened by determining a preset threshold to implement the process of determining the target POI type from the POI types. Specifically, for the preset threshold thre in step S404, j , can be set to the average value of TF-IDF of the warning area corresponding to all warning locations. If TF-IDF ij Score below thre j , indicating that the type of interest point i appearing around the warning location j is not important for the subsequent association analysis.

[0075] By TF-IDF ijBy calculating the correlation, the importance of different types of points of interest in the warning area corresponding to the location of the warning event can be obtained, that is, the types of points of interest can be screened according to the correlation. The prior art does not use this method to calculate the correlation. When a certain type of point of interest has a high probability of appearing in the warning area corresponding to the location of the occurrence, the point of interest type is considered important. However, in actual application, the probability of this type of point of interest appearing at the locations of all warning events is very high, which cannot explain the importance of this type of point of interest. For example, in the warning areas S(a), S(b), and S(c) corresponding to the locations a, b, and c of the warning events, the frequency of occurrence of points of interest of the type of restaurants and supermarkets is very high, which cannot explain that restaurants and supermarkets have a high correlation with the warning events corresponding to restaurants and supermarkets, nor can restaurants and supermarkets be used as the target point of interest types at S(a), S(b), and S(c). However, if at S(a) and S(b), the frequency of occurrence of POIs of restaurant and supermarket types is very low, and only the frequency of occurrence at S(c) is very high, then it can be considered that the two POI types of restaurant and supermarket are highly correlated with the warning event corresponding to S(c).

[0076] After determining the target POI type, the process of determining the POI type combination corresponding to the warning type in step S104 is as follows: Figure 6 As shown, it may further include:

[0077] Step S601: determining a target distribution matrix according to the distribution matrix and the target interest point type corresponding to each warning event;

[0078] Step S602: According to the target distribution matrix, based on the Apriori algorithm, determine the conditional probability value of the combination of different interest point types appearing in the warning type;

[0079] Step S603: According to the conditional probability value, determine the combination of interest point types corresponding to the warning type from the combinations of different interest point types.

[0080] The greater the conditional probability, the more times the combination of interest point types co-occurs in the warning type. Therefore, in step S603, the interest point type combination with the largest conditional probability can be taken as the interest point type combination most relevant to the current warning type.

[0081] In an optional embodiment, step S601 is as follows: Figure 7 As shown, it may further include:

[0082] Step S701: determining a Boolean matrix corresponding to the distribution matrix according to the target interest point type corresponding to each warning event;

[0083] Step S702: Use the Boolean matrix as the target distribution matrix; wherein the distribution vector corresponding to each warning event in the Boolean matrix satisfies: the value of the target interest point type corresponding to the warning event is 1, and the values ​​of the other interest point types are 0.

[0084] It should be noted that the Apriori algorithm (Association rule learning) is a method for rediscovering interesting relationships between variables in a large-scale data set. Its purpose is to use some interesting measures to identify strong rules found in the database. However, in the present invention, if only the Apriori algorithm is used, the occurrence is represented by a value of 1, and the non-appearance is represented by a value of 0, which cannot reflect the difference in the probability of occurrence of different interest point types, and the obtained results are not accurate. Therefore, the Apriori algorithm is selected to be used in conjunction with the TF-IDF algorithm. By using the number of interest points in the TF-IDF algorithm to screen the interest point type, the Boolean values ​​of different interest point types are differentiated, which makes up for the limitation of the Apriori algorithm that it cannot analyze dense matrices and effectively excludes most interest point type combinations.

[0085] Since the Apriori algorithm can only process Boolean matrices and cannot directly process numerical distribution matrices, in an optional embodiment, step S701 can be implemented in the following way: ij The interest point type with a score higher than threj is set to 1 in the distribution matrix, and for TF-IDF ij Score below thre j The type of interest point is set to 0 in the distribution matrix. In this way, the obtained distribution matrix can be a numerical matrix containing only the numbers 0 and 1, which just meets the setting standard of the Boolean matrix. No additional matrix conversion process is required to obtain the Boolean matrix, which greatly saves calculation time and improves calculation efficiency.

[0086] Furthermore, for the determination of interest point type combinations, a reduction idea can be adopted, that is, two laws are set in the Apriori algorithm:

[0087] (1) If a set is a frequent itemset, then all of its subsets are frequent itemsets.

[0088] (2) If a set is not a frequent itemset, then all its supersets are not frequent itemsets.

[0089] The frequent item set in the present invention refers to a combination of interest point types that often appear together. The above law is that after calculating the conditional probability value of each combination of interest point types appearing in the warning type, it is compared with the conditional probability value of the existing item set, and the frequent item set with the largest conditional probability is always retained.

[0090] For the Apriori algorithm, the specific process includes: first determine the frequent item set, then determine the probability value of each item in the frequent item set according to the preset association rules; finally, screen each item in the frequent item set step by step according to the preset probability value until a combination of interest point types is obtained. The following is a specific example to illustrate the process of determining frequent item sets. Assume that there are 5 first-level candidate item sets in the target distribution matrix, namely: [bread, milk], [bread, diapers, beer, eggs], [milk, diapers, beer, cola], [bread, milk, diapers, beer] and [bread, milk, diapers, cola]. First, obtain the number of occurrences of each combination of interest point types in the target distribution matrix (i.e., first-level candidate item sets), that is, milk 4 times, bread 4 times, diapers 4 times, beer 3 times, cola 2 times, and eggs 1 time; then screen the first-level candidate item set according to the preset number of times (for example, 3 times) to generate a second-level candidate item set. The candidate item sets are: [milk, bread], [milk, diapers], [milk, beer], [bread, diapers], [bread, beer], [diapers, beer]. The number of occurrences of each candidate item set in the secondary candidate item set is counted according to the first-level candidate item set, that is, [milk, bread] 3 times, [milk, diapers] 3 times, [milk, beer] 2 times, [bread, diapers] 3 times, [bread, beer] 2 times, [diapers, beer] 3 times. The secondary candidate item set is screened again according to the preset number of times (for example, 3 times) to generate the final tertiary candidate item set [milk, bread, diapers] 3 times, which is the frequent item set that needs to be determined.

[0091] The following is another specific example to explain in detail the specific role of the TF-IDF algorithm in the Apriori algorithm:

[0092] The known conditions are: one warning type has warning events at three locations {Sa, Sb, Sc}, assuming there are only three POI types {p1, p2, p3}, and the number of occurrences of the three POI types within a 1000-meter circular range around the three warning locations are: Sa: p1 = 20, p2 = 15, p3 = 15, Sb: p1 = 5, p2 = 20, p3 = 25, and Sc: p1 = 7, p2 = 3, p3 = 10;

[0093] In the absence of TF-IDF algorithm to filter the types of interest points, the following steps are included:

[0094] Step 1: Convert the POI type distribution quantity of each warning event location into Boolean type (0-1 value), Sa: p1 = 1, p2 = 1, p3 = 1, Sb: p1 = 1, p2 = 1, p3 = 1, Sc: p1 = 1, p2 = 1, p3 = 1; where p represents the distribution quantity;

[0095] Step 2: Calculate the probability of occurrence of POI type combinations with 1 combination item in all warning events, and obtain the candidate set C1: P(X={p1})=3 / 3=1, P(X={p2})=3 / 3=1, P(X={p3})=3 / 3=1; where P represents probability, X represents POI type combination variable, and P(X={p1}) represents the probability when variable X takes the value of {p1} combination;

[0096] Step 3: For the candidate set C1, remove the POI type combinations whose probability of occurrence is less than a given threshold, and obtain a frequent set L1 with 1 combination item. If L1 is an empty set, execute step 8.

[0097] Among them, if the given occurrence probability threshold is set to 1, all POI type combinations are eliminated; if the given occurrence probability threshold is set to less than 1, all POI type combinations are retained;

[0098] Step 4: For the frequent set L1, generate a POI type combination with 2 combination items through self-connection, calculate the occurrence probability of the POI type combination with 2 combination items in all warning events, and obtain the candidate set C2: P(X={p1,p2})=3 / 3=1, P(X={p1,p3})=3 / 3=1; P(X={p2,p3})=3 / 3=1;

[0099] Step 5: For the candidate set C2, remove the POI type combinations whose probability of occurrence is less than a given threshold, and obtain a frequent set L2 with 2 combination items. If L2 is an empty set, execute step 8.

[0100] Among them, if the given occurrence probability threshold is set to 1, all POI type combinations are eliminated; if the given occurrence probability threshold is set to less than 1, all POI type combinations are retained;

[0101] Step 6: For the frequent set L2, generate a POI type combination with 3 combination items through self-connection, calculate the occurrence probability of the POI type combination with 3 combination items in all warning events, and obtain the candidate set C3: P(X={p1,p2,p3})=3 / 3=1;

[0102] Step 7: For the candidate set C3, remove the POI type combinations whose probability of occurrence is less than a given threshold, and obtain a frequent set L3 with 3 combinations. If L3 is an empty set, execute step 8.

[0103] Among them, if the given occurrence probability threshold is set to 1, all POI type combinations are eliminated; if the given occurrence probability threshold is set to less than 1, all POI type combinations are retained;

[0104] Step 8: If the given occurrence probability threshold is set to 1, the final combination items are an empty set; if the given occurrence probability threshold is set to less than 1, the final combination items are {{p1},{p2},{p3},{p1,p2},{p1,p3},{p2,p3},{p1,p2,p3}}, and the algorithm ends.

[0105] It can be seen from the above method that without the TF-IDF algorithm to filter the POI types, the final POI type combination includes all three POI types, which have no difference between them, and the calculation result is not accurate. The present invention uses the TF-IDF algorithm to filter the POI types in advance, which specifically includes the following steps:

[0106] Step 1: Calculate the TF-IDF scores of the POI types of the three warning locations respectively:

[0107] TF-IDF(Sa,p1)=(20 / 50) / (log(3 / (3+1))+1)=0.5615470066530408

[0108] TF-IDF(Sa,p2)=(15 / 50) / (log(3 / (3+1))+1)=0.42116025498978055

[0109] TF-IDF(Sa,p3)=(15 / 50) / (log(3 / (3+1))+1)=0.42116025498978055

[0110] TF-IDF(Sb,p1)=(5 / 50) / (log(3 / (3+1))+1)=0.1403867516632602

[0111] TF-IDF(Sb,p2)=(20 / 50) / (log(3 / (3+1))+1)=0.5615470066530408

[0112] TF-IDF(Sb,p3)=(25 / 50) / (log(3 / (3+1))+1)=0.701933758316301

[0113] TF-IDF(Sc,p1)=(7 / 50) / (log(3 / (3+1))+1)=0.1965414523285643

[0114] TF-IDF(Sc,p2)=(3 / 50) / (log(3 / (3+1))+1)=0.08423205099795611

[0115] TF-IDF(Sc,p3)=(10 / 50) / (log(3 / (3+1))+1)=0.2807735033265204;

[0116] Step 2: For each POI type corresponding to the warning event, remove the POI type with a TF-IDF score lower than the mean, convert the POI type distribution quantity of each warning location into a Boolean type (0-1 value), and obtain Sa: p1 = 1, p2 = 0, p3 = 0, Sb: p1 = 0, p2 = 1, p3 = 1, Sc: p1 = 1, p2 = 0, p3 = 1;

[0117] Step 3: Calculate the probability of occurrence of POI type combinations with a combination number of 1 in all warning events, and obtain the candidate set C1: P(X={p1})=2 / 3=0.667, P(X={p2})=1 / 3=0.333, P(X={p3})=2 / 3=0.667;

[0118] Step 4: For the candidate set C1, remove the POI type combinations that are less than the given occurrence probability threshold, and obtain the frequent set L1 with 1 combination item. If L1 is an empty set, execute step 9;

[0119] Among them, if the given occurrence probability threshold is set to the mean of the occurrence probability, then L1={p1,p3};

[0120] Step 5: For the frequent set L1, generate a POI type combination with 2 combination items through self-connection, calculate the occurrence probability of the POI type combination with 2 combination items in all warning events, and obtain the candidate set C2: P(X={p1,p3})=1 / 3=0.333;

[0121] Step 6: For the candidate set C2, remove the POI type combinations that are less than the given occurrence probability threshold, and obtain the frequent set L2 with 2 combination items. If L2 is an empty set, execute step 9;

[0122] Wherein, if the given occurrence probability threshold is set to the mean of the occurrence probability, then L2 = {};

[0123] Step 7: For the frequent set L2, generate a POI type combination with 3 combination items through self-connection, calculate the occurrence probability of the POI type combination with 3 combination items in all warning events, and obtain the candidate set C3:

[0124] Step 8: For the candidate set C3, remove the POI type combinations that are less than the given occurrence probability threshold, and obtain the frequent set L3 with 3 combination items. If L3 is an empty set, jump to step 9;

[0125] Step 9: The final combination is {{p1},{p3},{p1,p3}}, and the algorithm execution ends.

[0126] Comparing the two execution results, we can see that using the TF-IDF algorithm to calculate the correlation before the Apriori algorithm can effectively screen the types of interest points, clearly reflecting the differences between different types of interest points, and obtaining a combination of interest point types that is more relevant to the warning type. At the same time, by comparing the correlation with the set threshold, the number of distributions with a value of 0 or 1 is obtained, which just meets the setting requirements of the Boolean matrix, without the need for redundant matrix conversion, and improves calculation efficiency.

[0127] In an optional embodiment, for the above process of filtering the candidate set C1 to obtain the frequent set L1, and then self-connecting the frequent set L1 to generate the candidate set C2, the number of loops varies depending on the number of items in the candidate set C1 and the filtering conditions. Therefore, in an embodiment of the present invention, it can be expressed as filtering a set of k-item interest point combination items to obtain k-item frequent sets, and then self-connecting the k-item frequent sets to obtain a (k+1)-item frequent set. After multiple filtering cycles, the interest point type combination is finally obtained.

[0128] Figure 8 A method for analyzing a warning event according to an embodiment of the present invention is shown, which specifically includes:

[0129] Step S801: receiving and storing warning events;

[0130] Step S802: determining a warning area corresponding to the occurrence location of the warning event, and obtaining the type of points of interest in the warning area and the number of points of interest under each type of point of interest;

[0131] Step S803: determining a distribution vector corresponding to each warning event according to the number of points of interest and the total number of types of points of interest;

[0132] Step S804: forming a distribution matrix corresponding to the warning event type according to the distribution vector;

[0133] Step S805: Based on the distribution matrix, the TF-IDF algorithm is used to determine the relevance between each point of interest type and each warning event;

[0134] Step S806: taking the interest point type whose correlation with the warning event is greater than a preset threshold as the target interest point type of the warning event;

[0135] Step S807: calculating the conditional probability of each interest point combination item in the k interest point combination item set in the same type of warning event data set;

[0136] Step S808: retaining the interest point combination items whose conditional probability is greater than the minimum conditional probability in the k-item set to obtain a k-item frequent set;

[0137] Step S809: self-connect the k frequent sets to obtain a (k+1) frequent set;

[0138] Step S810: Determine whether k+1 is less than the total number of combination items; if so, repeat steps S807 to S810; otherwise, continue to execute step S811: select the interest point type combination with the highest conditional probability as the interest point type combination corresponding to the warning type, and save the mapping relationship between the interest point type combinations corresponding to the warning types.

[0139] In an optional embodiment, after step S104, the following may be included: determining the warning strategy corresponding to the warning event belonging to the warning type according to the combination of interest points corresponding to the warning type. For example, after determining that the combination of interest points corresponding to the fire warning is gas stations, restaurants and supermarkets, fire prevention measures may be deployed at each gas station, restaurant and supermarket in the warning area, and fire-fighting items such as fire blankets, fire extinguishers and emergency water sources may be prepared in advance.

[0140] In an optional embodiment, the combination of target interest point types corresponding to the warning types can also be analyzed according to a preset time period. For example, at the end of each month, the warning events of the month are obtained, and the interest point type combination of the month is timely analyzed. Due to the rapid development of society, the speed of change of shops is relatively fast, and the interest point type in the warning area corresponding to the occurrence location will also change with the change of shops. Therefore, by regularly analyzing the interest point type combination according to the preset time period, better configuration of preventive measures can be achieved.

[0141] In an optional embodiment, if Fig. 9 As shown, after analyzing the warning events that have occurred, the method can also predict the warning events that have not occurred, specifically including:

[0142] Step S901: receiving an early warning query request, where the early warning query request indicates a location to be queried;

[0143] Step S902: determining one or more target interest points corresponding to the location to be generated;

[0144] Step S903: Determine and output the warning type of the warning event corresponding to the location to be occurred according to the point of interest type of one or more target points of interest.

[0145] Through the above analysis method, the co-occurrence pattern of the target point of interest type and the warning type can be realized. The point of interest type combination corresponding to the warning type can be analyzed, and then the warning strategy corresponding to the warning event belonging to the warning type can be obtained, so as to take corresponding preventive measures later; at the same time, it is also possible to make reasonable predictions for the location to be queried, thereby achieving a double prevention effect.

[0146] The analysis method of the warning event of the embodiment of the present invention analyzes the types of points of interest in the warning area corresponding to the warning event to obtain the target point of interest type with a high correlation with the warning event, and then analyzes the combination of target point of interest types according to the warning type to obtain the target point of interest type combination corresponding to each warning type, thereby associating the warning type with the target point of interest type combination to obtain the point of interest type combination with a higher correlation with the warning type. In this way, the most relevant point of interest type combination corresponding to the warning events belonging to the same warning type can be determined, so that according to the point of interest type combination of a certain region, it can be determined which types of warnings are needed in the region, without the need to perform single-point analysis for each warning event, thereby making the warning type and the target point of interest type combination associated with it transferable between different warning events, thereby improving the analysis efficiency of the warning event.

[0147] Furthermore, by using the TF-IDF algorithm to determine the target interest point type corresponding to the warning event, the analysis of irrelevant interest point types can be reduced, thereby further improving the analysis efficiency.

[0148] According to a second aspect of an embodiment of the present invention, there is provided an analysis device for warning events.

[0149] Fig.10 FIG. 1 is a schematic diagram of the main modules of the analysis device 1000 for warning events according to the second aspect of an embodiment of the present invention. Fig.10 As shown, including:

[0150] The acquisition module 1001 is used to acquire one or more warning events; the warning event indicates the occurrence location and warning type;

[0151] A determination module 1002 is used to determine a warning area corresponding to each occurrence location and one or more types of points of interest included in the warning area;

[0152] The analysis module 1003 is used for the warning events of the same warning type: analyzing the interest point types corresponding to the warning events to determine the target interest point types whose correlation with the warning events is greater than a preset threshold; and determining the interest point type combination corresponding to the warning type according to the target interest point types corresponding to the respective warning events in the same warning type.

[0153] In an optional embodiment of the present invention, the analysis module 1003 is also used to determine the number of points of interest of each point of interest type in the warning area corresponding to each of the warning events; determine the distribution vector corresponding to each of the warning events based on the number of points of interest and the total number of the types of points of interest; and determine the target point of interest type based on the distribution vector.

[0154] In an optional embodiment of the present invention, the analysis module 1003 is also used to obtain the distribution vector corresponding to each warning event in the same warning event type; based on the distribution vector, form a distribution matrix corresponding to the warning event type; based on the distribution matrix, use the TF-IDF algorithm to determine the correlation between each interest point type and each warning event; and use the interest point type whose correlation with the warning event is greater than a preset threshold as the target interest point type of the warning event.

[0155] In an optional embodiment of the present invention, the analysis module 1003 is also used to determine the occurrence probability of each interest point type in the warning area corresponding to the warning event according to the distribution vector corresponding to the warning event; determine the inverse document probability of each interest point type in the warning event type corresponding to the warning event according to the distribution matrix; and determine the correlation between each interest point type and the warning event based on the occurrence probability and the inverse document probability.

[0156] In an optional embodiment of the present invention, the analysis module 1003 is further used to determine a target distribution matrix based on the distribution matrix and the target interest point type corresponding to each of the warning events; based on the target distribution matrix, determine the conditional probability value of a combination of different interest point types appearing in the warning type based on the Apriori algorithm; based on the conditional probability value, determine the interest point type combination corresponding to the warning type from the combination of different interest point types.

[0157] In an optional embodiment of the present invention, the analysis module 1003 is also used to determine the Boolean matrix corresponding to the distribution matrix according to the target interest point type corresponding to each of the warning events; and use the Boolean matrix as the target distribution matrix; wherein the distribution vector corresponding to each warning event in the Boolean matrix satisfies: the value of the target interest point type corresponding to the warning event is 1, and the values ​​of the other interest point types are 0.

[0158] In an optional embodiment of the present invention, the determination module 1002 is also used to determine the warning area corresponding to the warning event with the coordinates of the occurrence location of the warning event as the center and a preset distance as the radius; and obtain the type of interest point in the warning area from the database.

[0159] In an optional embodiment of the present invention, the receiving module 1001 is also used to receive a warning query request, wherein the warning query request indicates the location to be queried; the determining module 1002 is also used to determine one or more target points of interest corresponding to the location to be occurred; the analyzing module 1003 is also used to determine and output the warning type of the warning event corresponding to the location to be occurred based on the point of interest types of the one or more target points of interest.

[0160] Fig.11 An exemplary system architecture 1100 is shown to which the method for processing a warning event or the apparatus for processing a warning event according to the embodiment of the present invention can be applied.

[0161] like Fig.11 As shown, system architecture 1100 may include terminal devices 1101, 1102, 1103, network 1104 and server 1105. Network 1104 is used between terminal devices 1101, 1102, 1103 and server 1105. Network 1104 may include various connection types, such as wired, wireless communication links or optical fiber cables, etc.

[0162] The user can use the terminal devices 1101, 1102, 1103 to interact with the server 1105 through the network 1104 to send an early warning query request or receive response information of the request (the early warning type of the early warning event corresponding to the location to be occurred), etc. Various communication client applications can be installed on the terminal devices 1101, 1102, 1103, such as online service applications, web browser applications, search applications, instant messaging tools, email clients, social platform software, etc.

[0163] The terminal devices 1101 , 1102 , and 1103 may be various electronic devices having a display screen and supporting web browsing, including but not limited to smart phones, tablet computers, laptop computers, and desktop computers, etc.

[0164] The server 1105 may be a server that provides various services, such as a background analysis server that provides support for analysis requests sent by users using terminal devices 1101, 1102, and 1103. The background analysis server may analyze and process the received warning query request and other data, and feed back the analysis results (such as the warning type of the warning event corresponding to the location to be occurred) to the terminal device.

[0165] It should be noted that the analysis method of the warning event provided in the first aspect of the embodiment of the present invention is generally executed by the server 1105 , and correspondingly, the analysis device of the warning event provided in the second aspect of the embodiment of the present invention is generally set in the server 1105 .

[0166] It should be understood that Fig.11 The number of terminal devices, networks and servers in the embodiment is only for illustration. Any number of terminal devices, networks and servers may be provided according to implementation requirements.

[0167] Reference below Fig.12 , which shows a schematic diagram of the structure of a computer system 1200 of a terminal device suitable for implementing an embodiment of the present invention. Fig.12 The terminal device shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present invention.

[0168] like Fig.12 As shown, the computer system 1200 includes a central processing unit (CPU) 1201, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 1202 or a program loaded from a storage part 1208 into a random access memory (RAM) 1203. In the RAM 1203, various programs and data required for the operation of the system 1200 are also stored. The CPU 701, the ROM 1202, and the RAM 1203 are connected to each other via a bus 1204. An input / output (I / O) interface 1205 is also connected to the bus 1204.

[0169] The following components are connected to the I / O interface 1205: an input section 1206 including a keyboard, a mouse, etc.; an output section 1205 including a cathode ray tube (CRT), a liquid crystal display (LCD), etc., and a speaker, etc.; a storage section 1208 including a hard disk, etc.; and a communication section 1209 including a network interface card such as a LAN card, a modem, etc. The communication section 1209 performs communication processing via a network such as the Internet. A drive 1210 is also connected to the I / O interface 1205 as needed. A removable medium 1211, such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, etc., is installed on the drive 1210 as needed, so that a computer program read therefrom is installed into the storage section 1208 as needed.

[0170] In particular, according to the embodiments disclosed in the present invention, the process described above with reference to the flowchart can be implemented as a computer software program. For example, the embodiments disclosed in the present invention include a computer program product, which includes a computer program carried on a computer-readable medium, and the computer program includes a program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from the network through the communication part 1209, and / or installed from the removable medium 1211. When the computer program is executed by the central processing unit (CPU) 1201, the above-mentioned functions defined in the system of the present invention are executed.

[0171] It should be noted that the computer-readable medium shown in the present invention may be a computer-readable signal medium or a computer-readable storage medium or any combination of the above two. The computer-readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, system or device, or any combination of the above. More specific examples of computer-readable storage media may include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In the present invention, a computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in combination with an instruction execution system, system or device. In the present invention, a computer-readable signal medium may include a data signal propagated in a baseband or as part of a carrier wave, which carries a computer-readable program code. This propagated data signal may take a variety of forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. Computer-readable signal media may also be any computer-readable medium other than computer-readable storage media that can send, propagate or transmit a program for use by or in conjunction with an instruction execution system, system or device. The program code contained on the computer-readable medium may be transmitted using any appropriate medium, including but not limited to: wireless, wire, optical cable, RF, etc., or any suitable combination of the above.

[0172] The flow chart and block diagram in the accompanying drawings illustrate the possible architecture, function and operation of the system, method and computer program product according to various embodiments of the present invention. In this regard, each box in the flow chart or block diagram can represent a module, a program segment, or a part of a code, and the above-mentioned module, program segment, or a part of a code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the box can also occur in a different order from the order marked in the accompanying drawings. For example, two boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram or flow chart, and the combination of the boxes in the block diagram or flow chart can be implemented with a dedicated hardware-based system that performs a specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.

[0173] The modules involved in the embodiments of the present invention may be implemented by software or hardware. The modules described may also be set in a processor, for example, it may be described as: a processor includes an acquisition module, a determination module and an analysis module. The names of these modules do not constitute a limitation on the modules themselves in some cases, for example, the acquisition module may also be described as a "module for acquiring one or more warning events".

[0174] As another aspect, the present invention further provides a computer-readable medium, which may be included in the device described in the above embodiment; or may exist independently without being assembled into the device. The above computer-readable medium carries one or more programs, and when the above one or more programs are executed by a device, the device includes:

[0175] Obtain one or more warning events; the warning event indicates the occurrence location and the warning type; determine the warning area corresponding to each occurrence location, and the type of one or more points of interest included in the warning area; for warning events of the same warning type: analyze the point of interest type corresponding to the warning event, and determine the target point of interest type whose correlation with the warning event is greater than a preset threshold; determine the point of interest type combination corresponding to the warning type based on the target point of interest types corresponding to each warning event in the same warning type.

[0176] The analysis method and device of the warning event of the embodiment of the present invention analyzes the types of points of interest in the warning area corresponding to the warning event to obtain the target point of interest type with a high correlation with the warning event, and then analyzes the combination of target point of interest types according to the warning type to obtain the target point of interest type combination corresponding to each warning type, thereby associating the warning type with the target point of interest type combination to obtain the point of interest type combination with a higher correlation with the warning type. Thus, the most relevant point of interest type combination corresponding to the warning event belonging to the same warning type can be determined, so that according to the point of interest type combination of a certain region, it can be determined which type of warning is required in the region, so that there is no need to perform a single point analysis for each warning event, so that the warning type and the target point of interest type combination associated with it have transferability between different warning events, thereby improving the analysis efficiency of the warning event. In addition, by using the TF-IDF algorithm to determine the target point of interest type corresponding to the warning event, the analysis of irrelevant point of interest types can be reduced, thereby further improving the analysis efficiency.

[0177] The above specific implementations do not constitute a limitation on the protection scope of the present invention. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions may occur depending on design requirements and other factors. Any modification, equivalent substitution and improvement made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A method for analyzing a warning event, characterized in that: The method comprises: Obtain one or more warning events; the warning event indicates the occurrence location and warning type; Determine a warning area corresponding to each of the occurrence locations, and one or more types of points of interest included in the warning area; For the warning events of the same warning type: Analyze the types of interest points corresponding to the warning event to determine target interest point types whose correlation with the warning event is greater than a preset threshold; Determine a point of interest type combination corresponding to the warning type according to the target point of interest types respectively corresponding to each warning event in the same warning type; The step of analyzing the types of interest points corresponding to the warning event to determine the target interest point types whose correlation with the warning event is greater than a preset threshold includes: Determine the number of points of interest of each point of interest type in the warning area corresponding to each of the warning events; Determining a distribution vector corresponding to each of the warning events according to the number of the points of interest and the total number of the types of the points of interest; The type of the target interest point is determined according to the distribution vector.

2. The method according to claim 1, characterized in that The determining the type of the target interest point according to the distribution vector comprises: Obtaining a distribution vector corresponding to each of the warning events in the same warning event type; According to the distribution vector, a distribution matrix corresponding to the warning event type is formed; Based on the distribution matrix, the TF-IDF algorithm is used to determine the relevance between each point of interest type and each of the warning events; The interest point type whose correlation with the warning event is greater than a preset threshold is used as the target interest point type of the warning event.

3. The method according to claim 2, characterized in that Determining the relevance between each point of interest type and each warning event using the TF-IDF algorithm based on the distribution matrix includes: Determine, according to the distribution vector corresponding to the warning event, the probability of occurrence of each point of interest type in the warning area corresponding to the warning event; Determine, according to the distribution matrix, the inverse document probability of each point of interest type in the warning event type corresponding to the warning event; Based on the occurrence probability and the inverse document probability, the degree of association between each point of interest type and the warning event is determined.

4. The method according to claim 2 or 3, characterized in that: The determining of a combination of interest point types corresponding to the warning type includes: Determine a target distribution matrix according to the distribution matrix and the target interest point type corresponding to each of the warning events; According to the target distribution matrix, based on the Apriori algorithm, determining the conditional probability value of the combination of different types of points of interest appearing in the warning type; According to the conditional probability value, a combination of interest point types corresponding to the warning type is determined from combinations of different interest point types.

5. The method according to claim 4, characterized in that Determining the target distribution matrix according to the distribution matrix and the target interest point type corresponding to each of the warning events includes: Determine a Boolean matrix corresponding to the distribution matrix according to the target interest point type corresponding to each of the warning events; Using the Boolean matrix as the target distribution matrix; The distribution vector corresponding to each warning event in the Boolean matrix satisfies that the value of the target interest point type corresponding to the warning event is 1, and the values ​​of the other interest point types are 0.

6. The method according to claim 1, characterized in that The determining of the warning area corresponding to each occurrence location and the type of one or more points of interest included in the warning area includes: Taking the coordinates of the occurrence location of the warning event as the center of the circle and the preset distance as the radius, determine the warning area corresponding to the warning event; The type of the point of interest in the warning area is obtained from a database.

7. The method according to claim 1, characterized in that Also includes: receiving an early warning query request, the early warning query request indicating a location where a query is to occur; Determine one or more target points of interest corresponding to the location to be occurred; According to the point of interest types of the one or more target points of interest, a warning type of the warning event corresponding to the location to be occurred is determined and output.

8. A device for processing early warning events, characterized in that: include: An acquisition module, used to acquire one or more warning events; The warning event indicates the occurrence location and warning type; A determination module, used to determine a warning area corresponding to each occurrence location, and one or more types of points of interest included in the warning area; An analysis module is used for the warning events of the same warning type: Analyze the types of interest points corresponding to the warning event to determine target interest point types whose correlation with the warning event is greater than a preset threshold; Determine a point of interest type combination corresponding to the warning type according to the target point of interest types respectively corresponding to each warning event in the same warning type; The analysis module is also used to: determine the number of points of interest under each point of interest type in the warning area corresponding to each of the warning events; determine the distribution vector corresponding to each of the warning events based on the number of points of interest and the total number of the point of interest types; and determine the target point of interest type based on the distribution vector.

9. An electronic device for analyzing warning events, characterized in that: include: one or more processors; a storage system for storing one or more programs, When the one or more programs are executed by the one or more processors, the one or more processors implement the method according to any one of claims 1 to 7.

10. A computer readable medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the method according to any one of claims 1 to 7 is implemented.

Images