WO2022034984A1

WO2022034984A1 - Device and method for predicting number of crime reports on basis of security and public data

Info

Publication number: WO2022034984A1
Application number: PCT/KR2020/017257
Authority: WO
Inventors: 김서연; 이민정; 이지윤; 목충협; 조용원; 황하은; 김성범
Original assignee: 고려대학교 산학협력단
Priority date: 2020-08-14
Filing date: 2020-11-30
Publication date: 2022-02-17

Abstract

Disclosed are a device and a method for generating a model for predicting the number of crime reports, on the basis of security and public data. The device for generating a prediction model comprises: a data collection unit for collecting data by using a crawling technique; a data processing unit for pre-processing the data; and a prediction model generation unit for generating a prediction model for predicting the number of crime reports at a prediction time point, by using the pre-processed data.

Description

Apparatus and method for predicting the number of crime reports based on security and public data

The present invention relates to a method for predicting the number of crime reports based on public security and public data, and in particular, a device for predicting the number of crime reports at the time of prediction by generating a predictive model using the number of crime reports and weather data, etc. and methods.

With the development of information and communication technologies such as artificial intelligence, the intelligence of security services is progressing, and major advanced countries such as the United States are making great efforts to effectively utilize the limited police force through police big data analysis on crimes over the past 10 years.

Recently, a system for preemptive response and investigation support is being developed by predicting dangerous situations in advance through crime terror detection through a crime prediction model, vehicle recognition of suspects, and tracking.

Although the number of police personnel is increasing to maintain social safety and security, there is a limit to just recruiting manpower, so the introduction of new technologies is urgently needed. Since the prior art performs crime prediction based on a traditional statistical method, it is difficult to secure prediction performance. Accordingly, the present invention intends to provide a result interpretation function by predicting the number of crimes based on a crime prediction model using machine learning technology and diagnosing predictive model factors.

The technical problem to be achieved by the present invention is to provide an apparatus and method for predicting the number of crime reports based on security and public data.

A predictive model generating apparatus according to an embodiment of the present invention includes a data collection unit that collects data using a crawling technique, a data processing unit that pre-processes the data, and the number of crime reports at the time of prediction using the pre-processed data. and a predictive model generator for generating a predictive model to predict.

The method for generating a predictive model according to an embodiment of the present invention is performed by a computing device including at least a processor, collecting data using a crawling technique, preprocessing the data, and using the preprocessed data and generating a predictive model for predicting the number of crime reports at the time of prediction.

A prediction apparatus according to an embodiment of the present invention includes a second data collection unit for collecting second data, a second data processing unit for pre-processing the second data, and any one of claims 5 to 7 It includes a prediction unit for predicting the number of crime reports at the time of prediction using a prediction model that is generated by the described prediction model generation method and that takes as an input the second data pre-processed.

According to the device and method for predicting the number of crime reports according to an embodiment of the present invention, it is possible to predict the number of crime reports at the time of prediction, and based on the prediction result, it is possible to efficiently process reports and crimes by jurisdiction. there is.

In order to more fully understand the drawings recited in the Detailed Description, a detailed description of each drawing is provided.

1 is a functional block diagram of an apparatus for generating a predictive model according to an embodiment of the present invention.

FIG. 2 shows important variables of crime occurrence predicted by the prediction unit shown in FIG. 1 .

3 is a functional block diagram of a prediction apparatus according to an embodiment of the present invention.

4 is a flowchart illustrating a prediction model generation method performed by the prediction model generation apparatus shown in FIG. 1 or a prediction method performed by the prediction apparatus illustrated in FIG. 3 .

Specific structural or functional descriptions of the embodiments according to the concept of the present invention disclosed herein are only exemplified for the purpose of explaining the embodiments according to the concept of the present invention, and the embodiment according to the concept of the present invention These may be embodied in various forms and are not limited to the embodiments described herein.

Since the embodiments according to the concept of the present invention may have various changes and may have various forms, the embodiments will be illustrated in the drawings and described in detail herein. However, this is not intended to limit the embodiments according to the concept of the present invention to specific disclosed forms, and includes all modifications, equivalents, or substitutes included in the spirit and scope of the present invention.

Terms such as first or second may be used to describe various elements, but the elements should not be limited by the terms. The above terms are used only for the purpose of distinguishing one component from another, for example without departing from the scope of the inventive concept, a first component may be termed a second component and similarly a second component A component may also be referred to as a first component.

When a component is referred to as being “connected” or “connected” to another component, it is understood that it may be directly connected or connected to the other component, but other components may exist in between. it should be On the other hand, when it is said that a certain element is "directly connected" or "directly connected" to another element, it should be understood that the other element does not exist in the middle. Other expressions describing the relationship between components, such as "between" and "immediately between" or "neighboring to" and "directly adjacent to", etc., should be interpreted similarly.

The terms used herein are used only to describe specific embodiments, and are not intended to limit the present invention. The singular expression includes the plural expression unless the context clearly dictates otherwise. In this specification, terms such as "comprise" or "have" are intended to designate that a feature, number, step, operation, component, part, or combination thereof described herein exists, but one or more other features It should be understood that it does not preclude the possibility of the presence or addition of numbers, steps, operations, components, parts, or combinations thereof.

Unless defined otherwise, all terms used herein, including technical or scientific terms, have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs. Terms such as those defined in a commonly used dictionary should be interpreted as having a meaning consistent with the meaning in the context of the related art, and should not be interpreted in an ideal or excessively formal meaning unless explicitly defined in the present specification. does not

Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings. However, the scope of the patent application is not limited or limited by these embodiments. Like reference numerals in each figure indicate like elements.

Referring to FIG. 1 , the predictive model generating apparatus 100 includes a data collecting unit 110 , a data processing unit 120 , a clustering unit 130 , a predictive model generating unit 140 , a predicting unit 150 , and storage. At least one of the units 160 may be included. That is, some components of the above-described predictive model generating apparatus 100 may be omitted according to embodiments, and the predictive model generating unit 140 may be implemented as a computing device including at least a processor and/or a memory.

The predictive model generating apparatus 100 uses the data collected by the data collection unit 110 or the data received through a predetermined input device and stored in the storage unit 160 to report the number of crimes (or the number of crimes) And/or it is possible to generate a predictive model for predicting crime occurrence factors. According to an embodiment, the predictive model generating apparatus 100 may predict the number of crime reports (or the number of crimes) and/or a crime occurrence factor by using the generated predictive model. In this case, the predictive model generating apparatus 100 may be referred to as a predictive apparatus.

The data collection unit 110 may collect necessary data in the process of generating a prediction model and performing a prediction operation. The data is data for a predetermined time period and may include at least one of first data and/or second data. The data collection unit 110 may collect data from a server that provides data through a crawling technique (or a portal site provided by the server). Data collected by the data collection unit 110 may be stored in the storage unit 160 .

The first data is data used in the learning process of the first predictive model for predicting the number of crime reports (or the number of crimes) and/or the process of predicting the number of reports of crimes (or the number of crimes) using the first predictive model. It may mean reported data. Table 1 below shows an example of crime report data.

월일month day	시간hour	신고 건수number of reports	……
4월 2일April 2	22~23시22-23 o'clock	44	……
4월 2일April 2	23~24시23-24 hours	22	……
4월 3일April 3	00~01시00 - 01:00	22	……
4월 3일April 3	01~02시01-2 o'clock	22	……
4월 3일April 3	02~03시02~03 o'clock	33	……
4월 3일April 3	03~04시03~04 o'clock	22	……
4월 3일April 3	04~05시04-05 o'clock	66	……
4월 3일April 3	05~06시05-06 o'clock	55	……
4월 3일April 3	06~07시06-07 o'clock	88	……
4월 3일April 3	07~08시07-8 o'clock	33	……
4월 3일April 3	08~09시08-09 o'clock	33	……
4월 3일April 3	09~10시09-10 o'clock	55	……
……	……	……	……

The first data may include at least one of a date, time, place, report content, and type of crime when a crime report occurred. The time at which the crime report occurred may mean the time at which the report was received or the time interval (time zone) at which the report was received. The time period may be divided into predetermined time units. For example, the time interval is a division of 24 hours a day into n (n is an integer) time unit, 0 to 1 o'clock, 1 to 2 o'clock, ... , may mean any one of 23 to 24 hours. When the time when the crime report occurs is defined as a time interval, the first data may include the number of reports received during the corresponding time interval. A place may mean a place where a crime occurred or a place where it was reported. According to an embodiment, the place is a coordinate that can identify the place (eg, GPS coordinates), an address (the address of the location where the complainant is located or the address of the location where the crime occurred), administrative unit, police station, district unit, jurisdiction of a police box, etc. can mean The report contents are the contents of the report of the crime reporter, and may mean audio data containing the contents of the report, text data containing the contents of the report, or summary data summarizing the contents of the report. The type of crime refers to the type of crime reported. It can mean any one of family, robbery/drugs, traffic, fraud, murder, women, suicide, theft, maintenance of order, youth, violence, customs, etc. there is. In the case of the Republic of Korea, crime report data corresponds to crime report data received through emergency phone number 112. The content included in the above-described first data may be a crawling result of the data collecting unit 110 , but at least a portion may be a result processed by the data processing unit 120 to be described later.

The second data may refer to data used in the learning process of the second predictive model for predicting at least one crime occurrence factor having the most importance among the crime occurrence factors and/or the crime occurrence factor prediction process using the second predictive model. can The second data may mean at least one of weather data, assembly data, public holiday data, local public data, and floating population data. The content included in the above-described second data may be a crawling result of the data collecting unit 110 , but at least some may be a result of processing by the data processing unit 120 to be described later.

The weather data may include past weather data and/or future weather forecast data. In addition, the weather data may include past weather data and/or weather forecast data for a certain area (province, city, district, etc.).

The assembly data may refer to assembly report data reported for the assembly, past assembly occurrence data, and the like, and may include information on the date and time of the assembly, the assembly place, the number of assembly people, and the like.

The holiday data may refer to data regarding legal holidays, temporary holidays, and the like.

Local public data includes the presence or absence of childcare facilities (daycare centers, kindergartens, etc.) in the area (eg, a pre-defined area (gu, competent police station, competent government office, etc.) Current status, incidence rate, frequency of occurrence, etc.), distribution of parks in the area, death statistics by cause of death in the area, presence of public sports facilities in the area, presence of leisure welfare facilities for the elderly in the area, status of persons with disabilities in the area (distribution rate of persons with disabilities) etc.), the current status of low-income single-parent families in the area (distribution rate, etc.), the classification of the area of use in the area, whether there is a national or public middle school in the area, whether or not apartment houses are built in the area, whether there are residential welfare facilities for the elderly in the area, It may mean at least one of the presence of medical institutions, the presence of national and public high schools in the region, the status of vacant houses in the region, and the resident registration population of the region.

The floating population data may mean the floating population at a specific point in time (or a specific time period) in the region, and may be provided by a telecommunication company, etc.

In some embodiments, the data collection unit 110 may be omitted from the predictive model generating apparatus 100 . In this case, the predictive model generating apparatus 100 may generate and/or predict a predictive model using data previously stored in the storage 160 .

The data processing unit 120 may pre-process the data collected by the data collection unit 110 and/or data previously received and stored in the storage unit 160 . Specifically, if a specific time is included in the crime report data included in the first data, it can be grouped by a predetermined time section and converted into crime report data including the time section and the number of reports in the corresponding time section. When the crime report data included in the first data includes a specific place (GPS coordinates, address, etc.), the data processing unit 120 transmits the crime report data to the competent police station, a predetermined corresponding area (gu, dong, etc.) , and/or the competent authority (jurisdictional district or police box) may be added or converted. If the first data includes specific report contents but does not include the type of crime, the type of crime may be determined based on the crime report and converted or added to crime report data including the crime type. To this end, the data processing unit 120 may convert voice data included in the crime report data into text data, and may determine the type of crime by extracting keywords described in the text data.

Also, the data processing unit 120 may delete data having missing values in the first data and/or the second data.

Data processed by the data processing unit 120 may be stored in the storage unit 160 . Also, when the preprocessing operation is unnecessary for the first data and/or the second data, the data processing unit 120 may be omitted from the predictive model generating apparatus 100 . In this case, the predictive model generating apparatus 100 may generate and/or predict a predictive model using data received by the data receiving unit 110 or data previously stored in the storage 160 . .

The clustering unit 130 may group the competent authorities into jurisdictions having a similar crime report pattern by using time-series crime report data.

Time-series crime report data may refer to first data. In this case, the first data may include the number of crime reports for each jurisdiction in a predetermined time unit (eg, 1 day or n hours).

In addition, the clustering unit 130 may measure the similarity (or distance) of the patterns of the time series data and group the jurisdictions according to the measured similarity (or distance), and a predetermined clustering algorithm may be used. An exemplary clustering algorithm is a K-means algorithm using dynamic time warping (DTW) as a distance scale.

The predictive model generating unit 140 reports a crime using the data received by the data receiving unit 110 , the data preprocessed by the data processing unit 120 , or data previously received and stored in the storage unit 160 . It is possible to create predictive models that predict the number and/or crime occurrence factors. The predictive model may include at least one of a first predictive model for predicting the number of crime reports using the first data and/or second data and a second predictive model for predicting a crime occurrence factor using the second data. .

Specifically, the predictive model generator 140 may generate the first predictive model by learning a time series model, for example, an artificial neural network such as a recurrent neural network (RNN) using the first data. Specifically, the prediction model generator 140 receives the number of crime reports before the prediction time using the first data, and learns the artificial neural network to predict the number of crime reports during a predetermined time interval (prediction interval) from the prediction time. By doing so, the first predictive model can be generated. Prediction of the number of crime reports may be performed up to k days (k is an integer) from the time the prediction operation is performed. For example, the first predictive model may predict the number of crime reports for 7 days from the time the prediction operation is performed by day. Table 2 below shows an example of predicting the number of crime reports by day from the time the prediction operation is performed until 7 days later.

예측 시점prediction time	예측 건수(범죄 신고 수)Number of forecasts (number of crime reports)
1일 후(내일)1 day later (tomorrow)	4949
2일 후2 days later	3434
3일 후3 days later	7979
4일 후4 days later	2323
5일 후5 days later	4444
6일 후after 6 days	6161
7일 후after 7 days	3939

According to an embodiment, the first prediction model may predict the number of crime reports for each predetermined time unit (eg, 6 hour units). As another example, the first prediction model predicts the number of crime reports for each unit, but may also predict the number of crime reports for each predetermined time unit with respect to at least a partial prediction time (eg, one day later). , the first prediction model may include prediction models for each of the clusters clustered by the clustering unit 130 . In this case, the predictive model generator 140 may train the predictive model for each of the clusters by using the learning data (data related to the corresponding cluster among the first data) corresponding to each of the clusters.

Also, the predictive model generator 140 may generate a second predictive model for determining a crime occurrence factor by using the second data. For example, the predictive model generator 140 uses the second data to learn a decision tree model (or boosting decision tree model) that extracts at least one important variable that has the greatest impact on crime occurrence, thereby making the second A predictive model can be created. Boosting decision tree model is a machine learning technique that combines multiple decision trees to output better results than a single decision tree. The decision tree is one of the supervised learning models that can perform both classification and regression. An exemplary boosting decision tree model is the CatBoost (Categorical Boost) technique.

The prediction unit 150 may perform a prediction operation using the prediction model generated by the prediction model generation unit 140 . Specifically, the prediction unit 150 may predict the number of crime reports by using the first predictive model and the second predictive model, and may additionally predict an important variable in the occurrence of a crime using the second predictive model.

When the number of crime reports is predicted using the first prediction model and the second prediction model, the prediction unit 150 uses the first data (eg, crime report data for the last 30 days) in a predetermined area (eg, jurisdiction It is possible to predict the number of crime reports by police station or competent authority). At this time, the number of crime reports in each zone is predicted using one first prediction model, or the number of crime reports in each zone is predicted using a prediction model specialized for each zone (that is, a prediction model generated using data of the cluster to which the zone belongs). The number of crime reports can be predicted. Also, the prediction unit 150 may predict the number of reported crimes during a predetermined prediction period by using the first data.

In the case of predicting an important variable in the occurrence of a crime using the second prediction model, the prediction unit 150 uses the first data and/or the second data to predict the importance of the occurrence of the crime at the time of prediction (eg, 3 days later). variable can be extracted. The prediction result of the important variable of the crime using the second prediction model is shown in FIG. 2 .

Efficient manpower allocation is possible through the prediction result of the prediction unit 150 . For example, it is possible to flexibly increase the number of police personnel at a time when the number of crime reports is expected to be high, or increase the number of police personnel in an area where the number of crime reports is expected to be high. In addition, it is possible to effectively prepare in advance according to an important variable in the occurrence of a crime. For example, if an important variable in the occurrence of crimes is 'childcare facilities', it will be possible to suppress the occurrence of crimes or to make early arrests by strengthening patrols and vigilance at childcare facilities.

In the storage unit 160, the data collected (or received) by the data collection unit 110, the data preprocessed by the data processing unit 120, the clustering result by the clustering unit 130, the predictive model generation unit ( The prediction model generated by 140 , the prediction result by the prediction unit 150 , and the like may be stored.

3 is a functional block diagram of a prediction apparatus according to an embodiment of the present invention. In the description of the prediction apparatus 300 illustrated in FIG. 3 , descriptions of contents overlapping with those described above will be omitted.

Referring to FIG. 3 , the prediction apparatus 300 may include at least one of a data collection unit 310 , a data processing unit 320 , a prediction unit 330 , and a storage unit 340 . That is, some components of the prediction apparatus 300 may be omitted according to embodiments, and the prediction apparatus 300 may be implemented as a computing device including at least a processor and/or a memory.

The prediction device 300 uses the data collected by the data collection unit 310 or the data received through a predetermined input device and stored in the storage unit 340 to determine the number of crime reports (or the number of crimes) and / Alternatively, the number of reported crimes (or the number of crimes) and/or factors of occurrence of crimes may be predicted by using a predictive model that predicts the factors causing the crime. The predictive model may refer to a predictive model generated by the predictive model generating apparatus 100 shown in FIG. 1 or a predictive model generating method performed by the predictive model generating apparatus 100 shown in FIG. 1 . .

The data collection unit 310 may collect necessary data in the process of performing the prediction operation. The data is data for a predetermined time period and may include at least one of first data and/or second data. Data collected by the data collection unit 310 may be stored in the storage unit 340 . The operation of the data collection unit 310 may be the same as that of the data collection unit 110 illustrated in FIG. 1 , and thus a detailed description thereof will be omitted.

In some embodiments, the data collection unit 310 may be omitted from the prediction apparatus 300 . In this case, the prediction apparatus 300 may perform a prediction operation using data previously stored in the storage unit 340 .

The data processing unit 320 may pre-process the data collected by the data collection unit 310 and/or data previously received and stored in the storage unit 340 . A detailed operation of the data processing unit 32 may be the same as that of the data processing unit 120 illustrated in FIG. 1 , and thus a detailed description thereof will be omitted.

Data processed by the data processing unit 320 may be stored in the storage unit 340 . Also, when the preprocessing operation is unnecessary for the first data and/or the second data, the data processing unit 320 may be omitted from the prediction apparatus 300 . In this case, the prediction apparatus 300 may perform a prediction operation using data received by the data receiver 310 or data previously stored in the storage 340 .

The prediction unit 330 may perform a prediction operation using the prediction model generated by the prediction model generation unit 140 shown in FIG. 1 . Specifically, the prediction unit 330 may predict the number of crime reports using the first prediction model, and may predict an important variable in the occurrence of a crime using the second prediction model. Since the detailed operation of the prediction unit 330 may be the same as that of the prediction unit 330 of FIG. 1 , a detailed description thereof will be omitted.

The storage unit 340 stores (or received) data collected by the data collection unit 310 , data preprocessed by the data processing unit 320 , a prediction model, and a prediction result by the prediction unit 330 . can be

FIG. 4 is a flowchart illustrating a prediction model generation method or a prediction method performed by the prediction model generation apparatus illustrated in FIG. 1 or the prediction apparatus illustrated in FIG. 3 .

First, data for generating a predictive model or data for performing a predictive operation are collected (S100). The data may include first data and/or second data. According to an embodiment, the data may be previously received and stored through a predetermined storage device, and in this case, step S100 may be omitted.

Received data or previously received and stored data is processed (or pre-processed) into a format for generating a predictive model or a format for performing a predictive operation (S200). Received data or previously received and stored data may be in a format for generating a predictive model or a format for performing a predictive operation, and in this case, step S200 may be omitted.

Thereafter, a plurality of competent authorities are grouped (or clustered) according to the pattern of the number of crime reports (S300). This is to generate different predictive models for each group, and when generating one predictive model, step S300 may be omitted. In addition, the prediction method other than the prediction model generation method may not include step S300.

Next, at least one predictive model may be generated ( S400 ). A detailed description of the predictive model generation method will be omitted.

Additionally, a prediction step may be performed (S500). In the prediction step, the number of crime reports (or the number of crimes) and/or important variables of the occurrence of crimes may be predicted using the generated prediction model.

The device described above may be implemented as a hardware component, a software component, and/or a set of hardware components and software components. For example, devices and components described in the embodiments may include, for example, a processor, a controller, an arithmetic logic unit (ALU), a digital signal processor, a microcomputer, a field programmable array (FPA), It may be implemented using one or more general purpose or special purpose computers, such as a Programmable Logic Unit (PLU), a microprocessor, or any other device capable of executing and responding to instructions. The processing device may execute an operating system (OS) and one or more software applications executed on the operating system. A processing device may also access, store, manipulate, process, and generate data in response to execution of the software. For convenience of understanding, although one processing device is sometimes described as being used, one of ordinary skill in the art will recognize that the processing device includes a plurality of processing elements and/or a plurality of types of processing elements. It can be seen that can include For example, the processing device may include a plurality of processors or one processor and one controller. Other Processing Configurations are also possible, such as a Parallel Processor.

The software may include a computer program, code, instructions, or a combination of one or more thereof, and configure the processing device to operate as desired or process it independently or in combination (Collectively) You can command the device. The software and/or data may be any kind of machine, component, physical device, virtual equipment, computer storage medium or device, to be interpreted by or provide instructions or data to the processing device. , or may be permanently or temporarily embodied in a transmitted signal wave (Signal Wave). The software may be distributed over networked computer systems and stored or executed in a distributed manner. Software and data may be stored in one or more computer-readable recording media.

The method according to the embodiment may be implemented in the form of program instructions that can be executed through various computer means and recorded in a computer-readable medium. The computer-readable medium may include program instructions, data files, data structures, etc. alone or in combination. The program instructions recorded on the medium may be specially designed and configured for the embodiment, or may be known and available to those skilled in the art of computer software. Examples of computer-readable recording media include magnetic media such as hard disks, floppy disks and magnetic tapes, optical recording media such as CD-ROMs and DVDs, and magnetic media such as floppy disks. - Includes hardware devices specially configured to store and execute program instructions, such as Magneto-optical Media, ROM, RAM, Flash memory, and the like. Examples of program instructions include not only machine language codes such as those generated by a compiler, but also high-level language codes that can be executed by a computer using an interpreter or the like. The hardware devices described above may be configured to operate as one or more software modules to perform the operations of the embodiments, and vice versa.

Although the present invention has been described with reference to the embodiment shown in the drawings, which is only exemplary, those skilled in the art will understand that various modifications and equivalent other embodiments are possible therefrom. For example, the described techniques are performed in a different order than the described method, and/or the described components of the system, structure, apparatus, circuit, etc. are combined or combined in a different form than the described method, or other components Or substituted or substituted by equivalents may achieve an appropriate result. Accordingly, the true technical protection scope of the present invention should be determined by the technical spirit of the appended claims.

Claims

a data collection unit that collects data using a crawling technique;

a data processing unit for pre-processing the data; and

A predictive model generating device including a predictive model generator for generating a predictive model for predicting the number of crime reports at a predictive time by using the pre-processed data.
According to claim 1,

The data includes crime report data that occurred during a predetermined time period,

The predictive model generation unit generates the predictive model by learning an artificial neural network using the crime report data,

Predictive model generation device.
3. The method of claim 2,

Further comprising a clustering unit that clusters the competent authorities according to the pattern similarity of the number of crime reports,

The predictive model generator generates a predictive model for each of the clusters,

Predictive model generation device.
4. The method of claim 3,

The clustering unit clusters using a K-means algorithm using DTW (Dynamic Time Warping) as a distance scale,

Predictive model generation device.
In the predictive model generation method performed by a computing device comprising at least a processor,

collecting data using a crawling technique;

pre-processing the data; and

A predictive model generating method comprising the step of generating a predictive model for predicting the number of crime reports at the time of prediction by using the pre-processed data.
6. The method of claim 5,

The data includes crime report data that occurred during a predetermined time period,

The generating of the predictive model comprises learning a Recurrent Neural Network (RNN) using the report data to generate the predictive model,

How to create a predictive model.
7. The method of claim 6,

After the pretreatment step,

Further comprising the step of clustering the competent authorities according to the pattern similarity of the number of crime reports,

The generating of the predictive model comprises generating a predictive model for each of the clusters,

How to create a predictive model.
a second data collection unit for collecting second data;

a second data processing unit pre-processing the second data; and

Claims 5 to 7, comprising a prediction unit that predicts the number of crime reports at the time of prediction using a prediction model that is generated by the prediction model generation method according to any one of claims 5 to 7 and uses the second data pre-processed as input prediction device.
9. The method of claim 8,

The second data includes crime report data that occurred during a predetermined time period,

prediction device.