TWI639963B - A system and a method for analyzing correlations of reported information - Google Patents

A system and a method for analyzing correlations of reported information Download PDF

Info

Publication number
TWI639963B
TWI639963B TW106146508A TW106146508A TWI639963B TW I639963 B TWI639963 B TW I639963B TW 106146508 A TW106146508 A TW 106146508A TW 106146508 A TW106146508 A TW 106146508A TW I639963 B TWI639963 B TW I639963B
Authority
TW
Taiwan
Prior art keywords
report
emotional
case
data
attribute
Prior art date
Application number
TW106146508A
Other languages
Chinese (zh)
Other versions
TW201931234A (en
Inventor
陳碧弘
朱陳彬
余憲全
陳韋金
賴彥如
Original Assignee
中華電信股份有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 中華電信股份有限公司 filed Critical 中華電信股份有限公司
Priority to TW106146508A priority Critical patent/TWI639963B/en
Application granted granted Critical
Publication of TWI639963B publication Critical patent/TWI639963B/en
Publication of TW201931234A publication Critical patent/TW201931234A/en

Links

Abstract

A reporting and correlation analysis system and method, comprising a report acceptance system, an Internet of Things sensor platform and an open data platform, connected to the platform and the emotional association analysis system of the report acceptance system, the emotional association analysis system includes Reporting sentiment collection module, emotional model generation module, emotional correlation analysis module and emotional related operation planning module, by investigating the reliability of the data source and the credibility of the emotional content, the report is converted. Calculating the support degree, the reliability and the degree of promotion between the plurality of attribute values, and mining the attribute values by analyzing the relationship between the selected attribute values in different cases by using multiple attribute values of the multiple interest characteristics. Potential association rules to provide a reference for managers.

Description

Reporting affair correlation analysis system and method
The invention relates to a report and relationship analysis system and method for a smart city, in particular to intelligently mining the association rules between the report and the situation, as a reference for the management department.
The service command center, the disaster relief command center or the citizen service line in the city has a special line for accepting people's reports. It is the most relied on the help of people when the safety of life and property is compromised. It accepts general criminal cases, traffic incidents and social order. , disaster events, services for the people, fires, emergency ambulances, consulting services, transfer services, grievance services, dispatch services, and other types of cases or services. At present, after the command centers receive the notification, the personnel on duty will input the content of the affair, and then according to the category of the disaster, the competent authority in charge of the business will send personnel to the scene for processing. In the event of false positives such as road name confusion, pronunciation errors, or drunken people, harassing calls, mental abnormalities, false reports, or unvoiced reports, it may result in wasteful allocation of public resources, such as invalid assignments or repeated assignments. The critically ill people, the golden time of missing reports, have a great impact on the follow-up ambulance work. If the reliability of the source of interest and the credibility of the content of the sentiment can be automatically determined in advance at the first time of notification, it will be effectively improved. The accuracy of case transfer and the effectiveness of support and assignment. When the case is processed, it is usually exchanged and passed on by the experience of the police or the ambulanceman to find out which areas of the jurisdiction are more likely to happen, what kind of behaviors in the jurisdiction are abnormal or people who often make troubles and troubles. If a large number of historical sentiment reports can be collected and aggregated and analyzed with the current sentiment notification, new rules of emotional association can be explored to prevent major losses by early prevention.
It can be seen that the above-mentioned conventional methods still have poor system availability, which is not a convenient and widely applicable design and needs to be improved.
In view of the shortcomings derived from the above-mentioned conventional methods, the inventor of the present invention has improved and innovated, and after years of painstaking research, he finally succeeded in researching and developing this documentary intelligence analysis system and method for a smart city.
The object of the present invention is to provide a report case correlation analysis system and a method thereof, which are to improve the intelligence analysis relationship of the smart city, use mathematical models and statistical analysis techniques, assist the government management department in decision-making, and dynamically adjust the protection strategy. To enhance the public's trust in smart cities.
The second objective of the present invention is to enhance the crisis handling ability of the smart city. By analyzing the attribute value of the report case, the government management department and the first-line duty personnel can immediately grasp the development status of the related cases and accelerate the situation. The speed of case handling makes the city operate more smoothly.
Another object of the present invention is to strengthen disaster prevention in a smart city. The contingency measures, through the analysis of the attribute association rules of the report case, can enable the government management departments and the public to prevent problems and improve the city's integrated disaster prevention capabilities.
A further object of the present invention is to analyze in advance the type of association between the informant and the reported case, which is summarized as one person reporting a situation or one person reporting multi-emotion, which can improve the reliability of the source of interest and the content of the situation. Reliability.
To achieve the above object, the present invention provides a report affair correlation analysis system, comprising: a report acceptance system configured to obtain a affair project as a data source for emotional association analysis; an Internet of Things sensor platform configured To provide IoT sensor data and obtain environmental sensing information; and an open data platform configured to provide open data; the emotional association analysis system includes: a report collection module, which is interposed in the report acceptance The system is configured to receive the emotional item, summarize the type of association between the informant and the emotional item, and determine the reliability of the data source and the credibility of the emotional content; the emotional model generating module, Connected to the Internet of Things sensor platform, configured to receive the information item from the report collection module, and obtain the environment sensing information from the Internet of Things sensor platform and from the open data platform The open information is used to convert the reported sentiment into multiple emotional characteristics; the emotional association analysis module is interfaced with the report receiving system, configured to receive from the The plurality of attribute features of the model generation module convert a plurality of attribute values of the plurality of interest features to analyze an association relationship of the selected attribute values in different cases; and an emotional association operation module configured Connecting to the emotional association analysis module, calculating support, reliability, and promotion between the plurality of attribute values, and mining potential associations between the attribute values Rules, as a basis for the judgment of the situation and the route planning of the patrol.
In the aforementioned report relationship analysis system, the sentiment-related operation planning module provides the analysis report as the data basis of the mathematical model and statistical analysis.
In the foregoing report relationship analysis system, the support degree calculated by the emotion-related operation and acquisition module is calculated by the probability that one of the plurality of attribute values intersects with other attribute values.
In the foregoing report relationship analysis system, the reliability calculated by the emotion-related operation module is calculated by calculating the probability that the support degree is one of the plurality of attribute values. .
In the foregoing report relationship analysis system, the degree of lift calculated by the sentiment-related operation module is calculated from the calculated probability of the trust in other attribute values.
The invention further provides a method for analyzing the relationship of report cases, which is to collect the report sentiment, collect the association type of the informant and the report case, and collect environmental sensing information and open data, and intelligently analyze the attribute value of the report. Correlation relationship, simplify the judgment of related cases, use statistical analysis to find out the potential association rules between case attributes, as the judgment of the situation of early warning, including: collecting the report and investing in the case, collecting the latest and historical report sentiment, Induct the type of association between the informant and the reported case, distinguish one person from the multi-emotion or one person to report the situation, calculate the reliability of the source of interest and the credibility level of the corresponding sentiment content; generate the emotional model steps, which are collected High-confidence sentiment, environmental sensing information and open data, classify data by name, order, interval or ratio metric category, replace deviation, error or invalid data with null value, establish eigenvalue set; analyze emotional association Step, based on Reporting sentiment, eigenvalue set and case attribute, generating a set of attribute values, analyzing whether there is a case in which the attribute value is associated; the operation-funding relationship-related step, which uses the report sentiment to calculate the support degree and reliability between the case attributes and Lifting degree, find out the potential association rules among the attributes, supplemented by the way of real-time dynamic presentation, providing managers as a reference for operations and decision-making.
In the foregoing report correlation analysis method, the step of collecting the report essay includes: the report essay collecting module immediately receives the latest report data of the informant transmitted by the report receiving system, and converts the report case into an affair project; the report case The intelligence collection module interfaces with the report acceptance system to obtain historical report data according to the informant and the specified time interval; the report collection module checks whether there is historical report data collected by the informant; if so, the notification The type of association between the applicant and the reported case is one person who informs the multi-emotion; if not, the type of association between the informant and the reported case is one person's notification of the situation; the reliability of the source of interest is calculated, and if the informant's effective number of notifications is 0, the reliability is 0. If the number of valid notifications of the informant is greater than 0, the reliability is the ratio of the number of valid notifications of the informant to the number of notifications of all the notified persons; and when the reliability of the source is 0 or The reliability is greater than the specified value, and the credibility level of the emotional content is high. The remaining levels are corresponding to the content of the emotional source. Reliability level.
In the foregoing report relationship analysis method, the step of generating the equity model includes: the sentiment model generating module obtains the information item of the report, and the information level of the credibility content is high; The resource generation module is connected to the Internet of Things sensor platform, and the sensing information provided by the Internet of Things sensor platform is collected according to the emotional project; the emotional model generation module is connected The open data platform is used to collect the open data provided by the open data platform according to the emotional project; the pre-processing of the emotional project data is performed, and the data is classified according to the name, order, interval or ratio metric, and the wrong data, invalid data or Offset value, replaced by null value; and establish a set of pending emotional feature values, where the feature value must be a valid data or null value.
In the foregoing report relationship analysis method, the analysis of the emotional association step includes: the emotional association analysis module is connected to the report acceptance system, and the report emotional characteristics generated by the emotional model generation module and the system predefined data are generated. The attribute and the range of attribute values are processed by the data conversion to make the metric categories of the same data attribute consistent; for the attribute value is null, the average or other fields are combined, divided or calculated to fill; according to the characteristics of the pending affair The set of values excludes the same case and establishes a set of attribute values of the case to be analyzed, wherein the attribute value must be a valid value; the emotional association analysis module selects a plurality of attribute values to be analyzed for the emotional relationship, and analyzes whether there are many a set of attribute values to be analyzed, including the plurality of attribute values; if there are multiple sets of case attribute values to be analyzed, the case is associated with the attribute value; and if there is no or only one set of case attribute values to be analyzed, Associate cases for non-attribute values.
In the foregoing report relationship analysis method, the operation and risk association step includes: the emotional association operation module obtains the set of attribute values of the case to be analyzed established by the emotional association analysis module, and includes the current processing and historical cases. Set; calculate the support between the attributes, that is, the probability that the attribute appears in the case at the same time; calculate the trust between the attributes, that is, the conditional probability that another attribute also appears after the occurrence of one attribute; Lifting degree, also That is, the degree of improvement of the probability of occurrence of another attribute after the occurrence of an attribute; and the reporting condition association rule of the summary item column and the presentation attribute value lifting degree greater than the specified threshold.
The aforementioned steps of collecting report affair are based on the informant and the report case received by the multiple report acceptance systems and the reporter's historical report, and the corresponding relationship between the informant and the report case is pre-instated to notify one person of one sentiment or one person's notice. More emotional. According to the notification time, geographical location, type of interest and the background environmental sound provided by the informant, the data pre-processing operation is carried out to generate the emotional project. Among them, the report case is extended to the arrival time of the automatic return of the duty officer or the ambulance personnel mobile device, the parking place of the vehicle and the record of the work record.
The steps of generating the emotional model are based on the emotional items collected by the report case, and are connected to the Internet of Things sensor platform to obtain images, speeds, etc. near the return position of the informant or the duty officer and the rescue personnel before the notification time occurs. Information, and then open the open data platform, obtain meteorological, traffic and other information at the time and place of the notification, and extract it into notification time, geographical location, type of interest, background environment sound type, suspect car key image, traffic speed, etc. Capital characteristics. After establishing the characteristics of the sentiment, the evaluation of the composition characteristics of the sentiment model is carried out. In the process, the results of the model of the emotional model are compared with the subsequent analysis of the relationship between the emotions, the expected analysis results and the cases encountered in practice. Whether the situation is consistent, and through the comparison of the emotional content of the different report receiving systems with the correlation model, the generated emotional model is confirmed to have robustness, appropriateness and good interpretation ability.
The aforementioned analysis of the emotional association step will generate the emotional characteristics of the emotional model, According to the pre-defined data attributes and attribute value ranges of the system, the data conversion processing job is performed, so that the attribute values of the same attribute are consistent, and the attribute value is ensured to be a valid value. The attribute is the option of association analysis. The important attribute helps to identify the relationship between the emotions, which can greatly improve the accuracy of the model. The system will prioritize the attributes with better analysis results, so that the effect of the correlation analysis can be The adjustment is getting better and better, and the analysis of the case report can be more intelligent.
The aforementioned operation and risk association step uses statistical analysis method to calculate the support degree, reliability and promotion degree between each attribute, and the system mines the potential association rules between attributes, and the real-time, dynamic and intuitive 3D stereogram and interactive analysis. Reporting method, presenting the association rules of time, region, meteorological and other attributes and case types, realizing instant query, analysis, presentation and release of the results of emotional association analysis, providing data basis for mathematical model and statistical analysis for smart city government management departments And can further use the emotional warning to assist decision-making and dynamically adjust the patrol guardian route.
The aforementioned report correlation analysis method can assist the police bureau command center system, the fire department disaster relief command center system, the citizen service system, the district service system and other report acceptance systems, and the informant information and the location of the case, the case description, etc. The data and the judgment of whether it is the result of the related situation, immediately presented on the screen of the operation center through the dashboard or webpage, constructing a convenient automatic transfer function, and solving the inconvenience that the duty officer can repeat the dictation or the operating system to report. It can shorten the timeliness of acceptance, transfer and assignment of information, speed up the assignment of police or ambulances, fire trucks to the scene, and give the public more peace of mind. This is the best system function that the city government attaches to the feelings of the people, so that the public can get the services of the modern science and technology command report system. Simultaneously With the factual description and location of the immediate rewards that have arrived at the on-site duty staff, and with the actual geographic image, the relevant authorities can more clearly and immediately grasp the actual situation of the situation, and then dispatch the support manpower to assist as needed.
Applying the aforementioned report sentiment correlation analysis method, it is possible to pre-discover which regions are in which time period, the frequency of occurrence of certain types of sentiments is particularly high, or if the traffic speed of any region suddenly slows down, the number of reports with a certain type of sentiment will increase rapidly. In addition to the fact that it can be aggregated into data and provide reference to the municipal management department, it can further analyze time, space, sound, video and other aspects, grasp the situation of neighboring environment and business type, and find out the true origin of the case more quickly. As soon as possible, the city will resume normal operation, effectively improving the timeliness and accuracy of personnel scheduling and assignment.
1‧‧‧Emotional Correlation Analysis System
11‧‧‧Reporting Information Collection Module
12‧‧‧Emotional Model Generation Module
13‧‧‧Emotional Correlation Analysis Module
14‧‧‧Emotional related operations module
2‧‧‧Report Acceptance System
3‧‧‧Internet of Things Sensor Platform
4‧‧‧Open Data Platform
S201-S207‧‧‧Steps
S301-S305‧‧‧Steps
S401-S406‧‧‧Steps
S501-S505‧‧‧Steps
The first figure is a schematic diagram of the architecture of the report correlation analysis system and method of the present invention.
FIG. 2 is a flow chart of collecting the case information of the report case correlation analysis system and method of the present invention.
FIG. 3 is a flow chart of generating a sentiment model of the report correlation analysis system and method of the present invention.
Figure 4 is a flow chart of the emotional association analysis of the report case correlation analysis system and method of the present invention.
FIG. 5 is a flow chart of the emotional connection operation of the report correlation analysis system and method of the present invention.
Please refer to FIG. 1 , which is a schematic diagram of the architecture of the report correlation analysis system and method of the present invention, which mainly includes: the emotional correlation analysis system 1 has the information collection of the report case, the generation of the emotional model, the analysis of the emotional relationship and The ability of intelligence-related operations. The sentiment association analysis system 1 includes a report situation collection module 11, a sentiment model generation module 12, an emotional association analysis module 13 and an emotional association operation module 14. The report collection module 11 receives the information of the informant, the notification time, the geographical location, the type of the situation, and the background environment sound of the report acceptance system, and summarizes the type of association between the informant and the report case. To determine the reliability of the source of the interest and the credibility of the content of the interest. The sentiment model generating module 12 obtains the environmental sensing information and the environment sensing information in the vicinity of the emotional occurrence time section and the geographical location of the emotional information item received by the report information collecting module 11 and the Internet of Things sensor platform. The open data platform obtains relevant open data related to the time zone and geographical location of the sentiment, and reconciles the report data into the notification time, geographical location, type of interest, the type of background environmental sound, the critical image of the suspect's car, and the speed of the traffic. A set of characteristics of the pending eigenvalues is established by the contingency feature. The sentiment association analysis module 13 combines the report sentiment features generated by the sentiment model generating module 12 according to the pre-defined data attributes and types of the system, and makes the attribute values of the same attribute consistent by the data conversion, and establishes a case to be analyzed. Corresponding time zone, case location, case type, background type, key image, vehicle speed range and other attribute values, analyze whether the selected plurality of attribute values exist in multiple cases at the same time, and find out the case in which the attribute values are related. The sentiment-related operation and acquisition module 14 uses a statistical analysis method to calculate support, reliability, and promotion among various attributes, and mining The potential association rules between attributes are presented in a dynamic and intuitive three-dimensional schema and interactive analysis report. The data base of mathematical model and statistical analysis is provided for the smart city government management department, and can be further used as the judgment of the emotional warning. The dynamic route planning of the patrol guard; the report acceptance system 2 has the ability to provide the content of the report, including a number of report acceptance systems. The report acceptance system 2 refers to the report acceptance system for emergency relief, ambulance or non-emergency citizen services in smart cities, including disaster relief command center, service command center, citizen Chen system, municipal mailbox, citizen hotline, smart district, etc. System, the above-mentioned citizen service system provides an interface for the report case collection module 11 to interface with the sentiment association analysis module 13, to receive or obtain instant and historical report content, in addition to the basic information of the report, including acceptance and transfer Relevant information such as assignment and processing results, as a source of case data for sentiment correlation analysis.
The Internet of Things Sensor Platform 3 is equipped with the ability to provide IoT sensor data. The main function is a device connection platform for accessing and sensing device data, and can receive intelligent environment sound type identifiers and intelligent video recording. Background Analysis of sensor data such as equipment and vehicle speed detectors. The Internet of Things sensor platform 3 is interfaced by the report model generation module 12, and can obtain environmental sensing information of a specified time range and region.
The Open Data Platform 4 is capable of providing open materials. The open materials refer to the information provided by the government or science and other fields that are screened and licensed. They are not restricted by the management mechanism and are allowed to be open to the public for free use and dissemination. At most, users are required to indicate the source of the data to everyone. For example: the open platform for meteorological data provided by the Central Meteorological Bureau of the Ministry of Communications Information, life safety and quality, transportation and communication, public information, leisure travel and other information provided by the county and city governments. The open data platform 4 is interfaced by the sentiment model generation module 12, and can obtain meteorological and traffic information in a specified time range and region.
Please refer to FIG. 2 , which is a flow chart of the report case collection system for the report case correlation analysis system and method of the present invention, which includes: Step S201: The report case collection module 11 immediately receives the latest report transmitted by the report acceptance system 2 . Report the information and convert the immediate report into an emotional project. The equity project is a report item collected at the time of receiving the notification, including the informant, notification time, geographical location, and type of interest. The informant may send the voice, video, newsletter or personally report the information to the public or the city's management department's designator or ambulanceman through manual notification or device; step S202: report case collection module 11 The report receiving system 2 obtains historical report data according to the informant and the specified time interval; step S203: the report sentiment collection module 11 checks whether there is historical report data collected by the informant; step S204: if yes, the notification The type of association between the applicant and the reported case is one person to inform the multi-emotion; step S205: If no, the type of association between the informant and the report case is one person to report a situation; step S206: calculating the reliability of the source of interest, if the notification If the number of valid notifications is 0, the reliability is 0. If the number of valid notifications of the informant is greater than 0, the reliability is the number of valid notifications of the informant and the number of notifications of all the notified persons. The ratio. The same case is included in the effective notification, but the consulting service is not included in the effective notification; Step S207: When the reliability of the source of interest is 0 or the reliability is greater than the specified value, the level of credibility of the content of the interest is high, and the remaining levels are based on the system. The user-defined stock source reliability ratio values respectively correspond to the credibility level of the interest content.
Please refer to FIG. 3, which is a flow chart of generating a sentiment model of the report relationship analysis system and method of the present invention, which includes: Step S301: The sentiment model generating module 12 obtains the report sentiment collection module 11 The affair content credibility level is high affair project data; step S302: the affair model generating module 12 is connected to the Internet of Things sensor platform 3, according to the affair item, including the notification time, the geographical location, and the type of the affair. Acquire sensing information received by sensors such as sound, video, and vehicle speed. For example: the type of background environment sound at the time of notification, the key image of the suspect's car near the geographical location, and the traffic speed near the geographic location. Some of the data may be error data, invalid data or deviation values; step S303: the equity model generation module 12 is connected to the open data platform 4, according to the emotional item, including the notification time, geographical location, and the type of interest, and the acquisition of the situation occurs. Open data of time and area meteorology (including temperature, rainfall, humidity) and traffic (average speed), some of which may have no value; Step S304: Perform pre-processing of the emotional project data, for different data, according to the metric category, It is classified into a nominal scale, an order scale, an interval scale, or a ratio scale. Wherein, the nominal scale belongs to a simple classification or name, and the comparison values may be equal or unequal; the order scale has a sequence, including The concept of size can compare the size of the value; the interval scale is the concept of order plus distance. The difference between the data can be calculated by adding and subtracting, and the value gap can be compared. The ratio scale has the concept of ratio and the absolute zero. The data can be multiplied and divided. , the numerical multiple or ratio can be calculated, according to the above metric category, the error data, the invalid data or the deviation value is judged, and replaced by the null value; step S305: establishing a set of the pending emotional feature values, wherein the feature value must be a valid data Or null value.
The reference example of the present invention is as follows: Suppose that S={S 1 , S 2 , . . . , S m } represents a set of m pens to be processed, and the set of feature values is respectively {a 1 , a 2 , .. ., a m } represents a set of m pen notification times, {b 1 , b 2 , ..., b m } represents a set of m pen geographic locations, {c 1 , c 2 , ..., c m } represents The set of m pen type, {d 1 , d 2 , ..., d m } represents the set of m pen background environment sound types, {e 1 , e 2 , ..., e m } indicates m pen suspicion The set of key images of people and vehicles, {f 1 , f 2 , ..., f m } represents the set of m-vehicle flow speeds, {g 1 , g 2 , ..., g m } represents the set of temperature and humidity of m pens {h 1 , h 2 , . . . , h m } represents a set of average speeds of the m pens, and the set of remaining feature values can be expressed in order.
Wherein, S i , S j , and S k respectively represent a set of three pending emotional feature values.
S i ={a i ,b i ,c i ,d i ,e i ,f i ,g i ,h i ,...} indicates that the notification content of the i-th report is: the interest feature a i indicates The notification time of the i-th report case is 2017-04-0404:30; the trait of b i indicates the geographical location of the i-th report, and its value is ○○○○○超商; The emotional characteristics c i represents the type of emotional information of the i-th report, the value of which is super-commercial robbery; the emotional characteristic d i represents the background environment sound type of the i-th report, and its value is the car whistle; The eligibility feature e i represents the key image of the suspect car of the i-th report, the value of which is close to the non-resident image file set near the informant; the emotional feature f i represents the traffic speed of the i-th report case, The value is 15 kilometers per hour; the emotional characteristics g i represents the temperature and humidity of the i-th report, the value is 24 ° C, 80%; the emotional characteristics h i represents the average speed of the i-th report, the value It is 25km/h.
Sj={a j ,b j ,c j ,d j ,e j ,f j ,g j ,h j ,...} indicates that the content of the report of the j-th report is: the character of interest a j indicates the first The notification time of j report case, its value is 2017-04-0501:25; the character of interest b j represents the geographical position of the j-th report, the value is the Ding Ding ○○ segment ○○ super business; The capital characteristic c j represents the type of sentiment of the j-th report, the value of which is a sudden burst of several gunshots; the emotional characteristic d j represents the background environment sound type of the j-th report, the value of which is Indoor environment; e-characteristics e j represents the key image of the suspect car of the j-th report, the value of which does not contain any valid data; the emotional characteristics f j represents the traffic speed of the j-th report, the value of which In order to not contain any valid information; the emotional characteristics g j represents the temperature and humidity of the j-th report, the value is 22.5 ° C, 85%; the emotional characteristics h j represents the average speed of the j-th report, and its The value is 35km/h.
S k ={a k ,b k ,c k ,d k ,e k ,f k ,g k ,h k ,...} denotes that the notification content of the k-th report is: the trait characteristic a k represents The notification time of the kth report is 2017-04-05 03:16; the emotional characteristics b k represents the geographical position of the kth report, and its value is the store of Wuji Road ○○; The capital characteristic c k represents the type of sentiment of the k-th report, and its value is the robbery; the emotional characteristic d k represents the background environment sound type of the k-th report, and its value is a noisy vocal; The feature e k represents the key image of the suspect car of the k-th report, the value of which is close to the suspect license plate image file set near the informant; the emotional feature f k represents the traffic speed of the k-th report case, and its value is 10 km per hour; the emotional characteristics g k represents the temperature and humidity of the k-th report, the value is 23.5 ° C, 82%; the emotional characteristics h k represents the average speed of the k-th report, the value is 40km /h.
Please refer to FIG. 4 , which is a flowchart of the emotional association analysis of the report relationship analysis system and method of the present invention, which includes: Step S401: The sentiment association analysis module 13 is connected to the report acceptance system 2, according to the situation. The characteristics of the report sent by the model generation module 12, such as: notification time, geographic location, type of interest, background environment sound type, suspect car key image, traffic speed, weather, traffic, etc., and system advance The defined data attribute and the attribute value range are processed by the data conversion processing, so that the data types of the same data attribute are consistent; step S402: for the attribute value is a null value, and the average value or other field combination, division or calculation is used to fill; Step S403: Excluding the same case equity according to the set of the emotional feature values to be processed, and establishing a set of attribute values of the case to be analyzed, wherein the attribute value must be a valid value, for example: time segment, case location, case type, background type , suspect, speed range, weather condition, comfort and other attributes; step S404: the emotional association analysis module 13 selects the analysis of the situation A plurality of attribute values associated analyze whether there are a plurality of cases to be analyzed attribute value set, which contains a plurality of attribute values.
Step S405: If there are multiple sets of case attribute values to be analyzed, The attribute value is associated with the case; step S406: if there is no or only one set of case attribute values to be analyzed, the non-attribute value is associated with the case.
The reference example is as follows: Assume that T={T 1 , T 2 ,..., T n } represents a set of n cases to be analyzed, and this set has excluded the same case, so n≦m. , the set of attribute values respectively denotes a set of n case receiving time segments by {A 1 , A 2 , ..., A n }, {B 1 , B 2 , ..., B n } represents n pens The set of case location at the place where the case occurred, {C 1 , C 2 ,..., C n } represents a set of n case types, {D 1 , D 2 ,..., D n } represents the background of n cases The set of types, {E 1 , E 2 ,..., E n } represents the set of suspects in n cases, {F 1 , F 2 ,..., F n } represents the range of nearby speeds in n cases The set, {G 1 , G 2 , ..., G n } represents a set of n weather conditions, {H 1 , H 2 , ..., H n } represents a set of n pen comforts, and the rest of the attribute values Collections can be represented in order. Where T p ={A p , B p , C p , D p , E p , F p , G p , H p ,...} indicates that the summary of the p-th case is: case attribute A p indicates The time zone of the p-th case is represented by a time zone consisting of morning, morning, noon, or night. The time zone of this example is early morning; the case attribute B p represents the location of the case of the p-th case, (latitude, The longitude range of the longitude indicates that the case of this example is (24.1483±N*0.0001°, 120.6730°±N*0.0001°); the case attribute C p indicates the type of case in the p case, the case of this case The type is ordinary robbing; the case attribute D p indicates the background type of the p-th case, the background type of this case is super-business; the case attribute E p represents the suspect's collection of the p-th case, and the suspect in this case is the place where the affair occurs. Nearby, the critical image of the suspect's car is analyzed differently than the suspicious person; the case attribute F p indicates the range of the speed of the p-th case, which is expressed by less than the average speed, and the speed range is determined to be less than or equal to the occurrence of the situation. The average speed of the time near the location, the speed range of this example is small 20 km to 60 km; case property G p represents weather conditions the p pen case, the present embodiment of the weather condition is Clear; case attribute H p represents the comfort of p pen case, the present embodiment of comfort for comfort to hot .
In this embodiment, the four attributes of the suspect, the time segment, the case location and the case type are selected as the attributes of the association analysis, and the suspect E w is ○○○, the time segment A x is the early morning, and the case type B y is the ordinary robbing. And the background type C z is the relationship of the four attribute values of the super quotient, and finds the E W = ○ ○ ○ from the set T={T 1 , T 2 , ..., T n } of the case to be analyzed. a x = morning and B y = normal snatch and C z = background type of case, if three items case of T p is present, T q and T r is met, T p, T q and T r three pen case for suspect The time zone, case type, and background type are cases in which attribute values are associated.
The set of cases to be analyzed T={T 1 , T 2 , . . . , T n } is obtained by inductive analysis from the set of pending emotions S={S 1 , S 2 , . . . , S m } The attribute value associated cases T p , T q and T r are arranged in chronological order, and it can be found that the cases of T p , T q and T r are related cases of multiple cases of suspects.
Please refer to FIG. 5 , which is a flowchart of the emotional connection operation of the report relationship analysis system and method of the present invention. From the relationship between a series of attributes, through the support degree, the reliability and the promotion degree, find out the attributes between the attributes. The potential association rule, the action-related management module implementation step is: Step S501: The sentiment-related operation and acquisition module 14 obtains the set of case attribute values to be analyzed established by the sentiment association analysis module 13. The collection includes the current processing and historical cases, and is used for the subsequent use of the statistical analysis method to find out the data preparation of the potential association rules between the attributes; step S502: calculating the support degree between the attributes, that is, the probability that the attribute appears in the case at the same time. For example, the case attribute B p indicates the case location of the p case, the case attribute C p indicates the case type of the p case, the case attribute B p and the case attribute C p appear in the case at the same time, and the support degree calculation formula P(Bp ∩ Cp); Step S503: Calculating the reliability between the attributes, that is, the conditional probability that another attribute also appears after an attribute appears, for example, the case attribute G p represents the weather condition of the p-th case After the occurrence of the case attribute G p , the conditional probability of the case attribute C p also appears, and the reliability calculation formula is Step S504: Calculate the degree of promotion between the attributes, that is, the degree of improvement of the probability of occurrence of another attribute after the occurrence of an attribute; for example, after the case type C p occurs first, the probability of occurrence of the suspect E p is The degree of improvement, the formula for calculating the degree of lift is Step S505: The summary column and the report attribute association rule whose presentation attribute value lifting degree is greater than the specified threshold.
The reference examples are as follows: Assume that the probability of the occurrence of the case attribute value is the case location B p (24.1483 ° ± 0.0001 °, 120.6730 ° ± 0.0001 °), the probability of occurrence is 50%; the case type C p is the mass fight, the probability of occurrence 20%; the probability of occurrence of case location B p and case type C p is 40%. If the minimum support is set to 30% and the minimum reliability is set to 60%, the support, reliability and promotion are calculated separately: Support: P(Bp ∩ Cp)=40%; Reliability: Lifting degree: The resulting association rule is that 40% of all cases occur in areas with latitude and longitude (24.1483±±0.0001°, 120.6730°±0.0001°) and the case type is mass fight, with 40% support exceeding the support threshold. 30%; and in the area where the latitude and longitude is (24.1483±±0.0001°, 120.6730°±0.0001°), 80% of the emotions are for the masses, and the reliability is 80% exceeding the trust threshold of 60%; The association rule has a degree of accuracy improvement of 4 compared to the random guess. The rule found in this case will be a reference rule with reference value, which can be expressed as follows: {case location = (24.1483 ° ± 0.0001 °, 120.6730 ° ± 0.0001 °)} -> {case type = mass fight}
When the degree of support is higher, the higher the degree of significance among the representative attributes, the more worthy of further discussion on behalf of the attribute combination; the higher the reliability, the higher the accuracy between the representative attributes, the more reference value; when the degree of promotion is greater than 1, the association The rules have a reference meaning. The invention can find out the types of cases that often occur from the case area, and can also use the methods of support, reliability and promotion to further find out whether the case type is related to the time of the incident, and use the case hot zone and the timeline presentation mode. It will be smart and flexible to patrol the police.
The report and emotion correlation analysis system and method provided by the invention have The following advantages: the report collection module is to summarize the type of association between the informant and the report case as one person to report multiple interest or one person to report a situation, calculate the reliability of the source of interest and the level of credibility of the content, as a dispatch A reference to prioritization and whether it is included in the conditions of the association analysis.
The sentiment model generation module adds the report sentiment to the sentiment-related sensing information and open data, generates a set of pending emotional feature values, classifies according to the metric category, and excludes the deviation value, error or invalid data, so as to make the situation The efficiency of the model generation is significantly improved.
The affair correlation analysis module can improve the association value of the attribute value of the report case, reduce the associated cases and assign different contract windows, and analyze the implicated cases with the attribute values, which is highly practical.
The Emotional Related Operations Module can effectively assist the smart city government management department to obtain sufficient information to formulate countermeasures so that the damage caused by similar mistakes can be minimized.
The detailed description of the present invention is intended to be illustrative of a preferred embodiment of the invention, and is not intended to limit the scope of the invention. The patent scope of this case.
To sum up, this case is not only innovative in terms of technical thinking, but also has many of the above-mentioned functions that are not in the traditional methods of the past. It has fully complied with the statutory invention patent requirements of novelty and progressiveness, and applied for it according to law. Approved this invention patent application, in order to invent invention, to the sense of virtue.

Claims (10)

  1. A report sentiment correlation analysis system, comprising: a report acceptance system configured to obtain a sentiment project as a source of information for emotional association analysis; an Internet of Things sensor platform configured to provide IoT sensor data And obtaining environmental sensing information; an open data platform configured to provide open data; an emotional association analysis system comprising: a report collection module, which is interfaced to the report acceptance system, configured to receive the situation The type of association between the funded project, the inductive informant and the emotional project, and the reliability of the source of the data and the credibility of the content of the information; the sentiment model generating module, which is connected to the Internet of Things sensor a platform configured to receive the benefit item from the report collection module and obtain the environmental sensing information from the Internet of Things sensor platform and the open data from the open data platform to convert the The report sentiment is a plurality of emotional characteristics; the sentiment correlation analysis module is interposed in the report receiving system, configured to receive the plurality of situations from the sentiment model generating module Feature, converting a plurality of attribute values of the plurality of affiliation features to analyze an association relationship of the selected attribute values in different cases; and an affair-related operation planning module configured to be connected to the affiliation association analysis module, Calculate the support, trust, and lift between the multiple attribute values, and mine the potential association rules between the attribute values. For the judgment of the emotional warning and the route planning of the patrol.
  2. According to the report sentiment correlation analysis system described in Item 1 of the patent application scope, the relationship-related operation planning module provides the analysis report as the data basis of the mathematical model and statistical analysis.
  3. According to the report case correlation analysis system described in claim 1 of the patent application scope, the support degree calculated by the emotion-related operation and acquisition module is intersected by one of the plurality of attribute values and other attribute values. The probability is calculated.
  4. According to the report sentiment correlation analysis system described in claim 3, wherein the reliability calculated by the emotional associated operation module is calculated by the supported degree in the plurality of attribute values. One of the odds is calculated.
  5. According to the report sentiment correlation analysis system described in claim 4, wherein the upgrade degree calculated by the emotional associated operation module is calculated by the calculated reliability in other attribute values. The probability is calculated.
  6. A report correlation analysis method for reporting cases, which collects the report sentiment, collects the type of association between the informant and the report case, collects environmental sensing information and open data, intelligently analyzes the relationship between the report attribute values, and simplifies the association. Judging the case, using statistical analysis to find out the potential association rules between the case attributes, as the judgment of the situation of the situation, including: collecting the information on the case report, collecting the latest and historical report, summarizing the informant and reporting The type of association of sentiment, distinguishing one person from multiple interest or one person to report a situation, calculating the reliability of the source of interest and The level of credibility of the content of the sentiment; the step of generating the emotional model, which combines high-confidence sentiment, environmental sensing information and open data, classifying the data by name, order, interval or ratio metric, empty The value replaces the deviation value, the error or the invalid data, establishes the feature value set; analyzes the emotional association step, which generates the attribute value set according to the report case, the feature value set and the case attribute, and analyzes whether there is a case in which the attribute value is associated; The operation and risk association step is to use the report case to calculate the support degree, reliability and promotion between the case attributes, find out the potential association rules between the attributes, and provide the managers with the way of real-time dynamic presentation. Reference for operations and decision making.
  7. For example, the method for analyzing the sentiment relating to the report described in claim 6 wherein the step of collecting the report includes: the report collection module immediately receives the latest report data of the informant transmitted by the report acceptance system, and will report the case. The funds are converted into an emotional project; the report collection module is connected to the report acceptance system, and the historical report data is obtained according to the informant and the specified time interval; the report collection module checks whether the reporter has been collected. Historical report data; if so, the type of association between the informant and the reported case is one person to inform the multi-emotion; if not, the type of association between the informant and the report case is one person's notification; Calculate the reliability of the source of interest. If the number of valid notifications of the informant is 0, the reliability is 0. If the number of valid notifications of the informant is greater than 0, the reliability is the number of valid notifications of the informant, and all the notified persons have notified. The ratio of the number; and when the reliability of the source is 0 or the reliability is greater than the specified value, the level of credibility of the content of the interest is high, and the remaining levels are respectively corresponding to the reliability ratio of the source of the information customized by the system user. The level of credibility of the content of the affair.
  8. For example, the method for analyzing the relationship of interest according to claim 6 of the patent application scope, wherein the step of generating the emotional model comprises: the emotional model generating module obtains the reporting sentiment collection module, and the level of credibility of the emotional content is The high-quality project data; the emotional model generating module is connected to the Internet of Things sensor platform, and the sensing information provided by the Internet of Things sensor platform is collected according to the emotional project; the emotional model generating module is connected An open data platform that collects open data provided by the open data platform according to the emotional project; conducts pre-processing of the emotional project data, classifies the data according to the name, order, interval or ratio metric, and judges the wrong data, invalid data or deviation The value is replaced by a null value; and a set of pending eigenvalue values is established, wherein the eigenvalue must be a valid data or null value.
  9. For example, the method for analyzing the sentiment of the case described in claim 6 of the patent application scope, wherein the step of analyzing the emotional relationship includes: The affair correlation analysis module interfaces with the report acceptance system, according to the report traits generated by the stimuli model generation module and the pre-defined data attributes and attribute value ranges of the system, and the data conversion processing operation enables the measurement of the same data attribute. The categories are consistent; for the attribute value is null, it is filled by the combination or division or calculation of the average value or other fields; according to the set of the eigenvalues to be processed, the same case is excluded, and the set of the attribute values of the case to be analyzed is established. The attribute value must be a valid value; the emotional association analysis module selects a plurality of attribute values to be analyzed for the emotional association, and analyzes whether there are multiple sets of attribute values to be analyzed, and includes the plurality of attribute values; If the set of attribute values of the case to be analyzed is the case of the attribute value association; and if there is no or only one set of case attribute values to be analyzed, the case is a non-attribute value.
  10. For example, the method for correlating the sentiment of the case according to claim 9 of the patent application scope, wherein the step of correlating the situation of the operation includes: the emotional related operation planning module obtains the set of attribute values of the case to be analyzed established by the emotional association analysis module , including the current collection of processing and historical cases; calculating the support between the attributes, that is, the probability that the attribute appears in the case at the same time; calculating the reliability between the attributes, that is, after the occurrence of an attribute, The conditional probability that another attribute also appears; the degree of promotion between the attributes is calculated, that is, the degree of improvement of the probability of occurrence of another attribute after the occurrence of one attribute; and the degree of lifting of the summary column and the presentation attribute value is greater than the specified threshold. Reporting rules for the association of sentiment.
TW106146508A 2017-12-29 2017-12-29 A system and a method for analyzing correlations of reported information TWI639963B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
TW106146508A TWI639963B (en) 2017-12-29 2017-12-29 A system and a method for analyzing correlations of reported information

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
TW106146508A TWI639963B (en) 2017-12-29 2017-12-29 A system and a method for analyzing correlations of reported information

Publications (2)

Publication Number Publication Date
TWI639963B true TWI639963B (en) 2018-11-01
TW201931234A TW201931234A (en) 2019-08-01

Family

ID=65034205

Family Applications (1)

Application Number Title Priority Date Filing Date
TW106146508A TWI639963B (en) 2017-12-29 2017-12-29 A system and a method for analyzing correlations of reported information

Country Status (1)

Country Link
TW (1) TWI639963B (en)

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20130268196A1 (en) * 2012-04-05 2013-10-10 Quang Binh Dam System and method of automated acquisition, correlation and display of power distribution grid operational parameters and weather events
US20140283083A1 (en) * 2013-03-15 2014-09-18 Tenable Network Security, Inc. System and method for correlating log data to discover network vulnerabilities and assets
TWM517758U (en) * 2015-10-08 2016-02-21 Univ China Technology Bridge safety detection device with IoT
TW201642677A (en) * 2015-05-18 2016-12-01 Univ Southern Taiwan Sci & Tec System and method for dynamic management of context awareness service to multiple users

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20130268196A1 (en) * 2012-04-05 2013-10-10 Quang Binh Dam System and method of automated acquisition, correlation and display of power distribution grid operational parameters and weather events
US20140283083A1 (en) * 2013-03-15 2014-09-18 Tenable Network Security, Inc. System and method for correlating log data to discover network vulnerabilities and assets
TW201642677A (en) * 2015-05-18 2016-12-01 Univ Southern Taiwan Sci & Tec System and method for dynamic management of context awareness service to multiple users
TWM517758U (en) * 2015-10-08 2016-02-21 Univ China Technology Bridge safety detection device with IoT

Also Published As

Publication number Publication date
TW201931234A (en) 2019-08-01

Similar Documents

Publication Publication Date Title
US20170140468A1 (en) Vehicle router
US20190278768A1 (en) Qa based on context aware, real-time information from mobile devices
US9307385B2 (en) Utilizing information about mobile communication devices with respect to an area of interest
Ludwig et al. CrowdMonitor: Mobile crowd sensing for assessing physical and digital activities of citizens during emergencies
US20170293847A1 (en) Crime risk forecasting
US10217169B2 (en) Computer system for determining geographic-location associated conditions
US20180053401A1 (en) Predictive analytics for emergency detection and response management
Thompson et al. Evacuation from natural disasters: a systematic review of the literature
US9805311B1 (en) Event forecasting system
US10178537B2 (en) Emergency messaging system and method of responding to an emergency
US10504357B2 (en) Apparatus and methods for distributing and displaying emergency communications
US10484856B1 (en) Normalizing insufficient signals based on additional information
US9867023B2 (en) Emergency messaging system and method of responding to an emergency
US8923799B2 (en) Method and system for an automated dispatch protocol
CA2972451C (en) Method and apparatus for prediction of a destination and movement of a person of interest
Hirschfield et al. Mapping and analysing crime data: Lessons from research and practice
Maltz et al. Mapping crime in its community setting: Event geography analysis
Kennedy et al. Risk clusters, hotspots, and spatial intelligence: risk terrain modeling as an algorithm for police resource allocation strategies
US8041743B2 (en) Systems and methods for providing semantically enhanced identity management
US9852342B2 (en) Surveillance system
Southworth Regional evacuation modeling: A state-of-the-art review
Burton et al. National victims of crime survey South Africa 2003
US7411493B2 (en) User-centric event reporting
US20150009047A1 (en) Method and apparatus for vehicle parking spaces management using image processing
US10559180B2 (en) Building risk analysis system with dynamic modification of asset-threat weights