CN111768031B - Method for predicting crowd gathering trend based on ARMA algorithm - Google Patents
Method for predicting crowd gathering trend based on ARMA algorithm Download PDFInfo
- Publication number
- CN111768031B CN111768031B CN202010588491.1A CN202010588491A CN111768031B CN 111768031 B CN111768031 B CN 111768031B CN 202010588491 A CN202010588491 A CN 202010588491A CN 111768031 B CN111768031 B CN 111768031B
- Authority
- CN
- China
- Prior art keywords
- predicted
- value
- day
- aggregate
- daily
- Prior art date
- Legal status (The legal status 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 status listed.)
- Active
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/29—Geographical information databases
Landscapes
- Engineering & Computer Science (AREA)
- Business, Economics & Management (AREA)
- Theoretical Computer Science (AREA)
- Human Resources & Organizations (AREA)
- Databases & Information Systems (AREA)
- Economics (AREA)
- Strategic Management (AREA)
- General Physics & Mathematics (AREA)
- Physics & Mathematics (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- Tourism & Hospitality (AREA)
- Marketing (AREA)
- General Business, Economics & Management (AREA)
- Entrepreneurship & Innovation (AREA)
- Development Economics (AREA)
- Game Theory and Decision Science (AREA)
- Remote Sensing (AREA)
- Data Mining & Analysis (AREA)
- General Engineering & Computer Science (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses a method for predicting crowd gathering trend based on ARMA algorithm, which aims to find a more effective implementation scheme for crowd gathering trend prediction, and comprises the following steps: acquiring daily aggregate value and daily activity information in p days before the day to be predicted of the preset place; obtaining a predicted aggregate value of the day to be predicted based on the activity information of the day and the daily aggregate value and the daily activity information in p days; and acquiring a daily aggregate numerical value in r days before the day to be predicted, and confirming whether the predicted aggregate numerical value is abnormal or not based on the daily aggregate numerical value in r days and the predicted aggregate numerical value, if so, displaying that the predicted aggregate numerical value is abnormal. The invention adopts the long-time multidimensional crowd gathering data of attention, utilizes ARMA algorithm to extract the potential association between the event which cannot be judged manually and the time, and overcomes the defect of experience judgment on the crowd gathering degree in the traditional work.
Description
Technical Field
The invention relates to the field of big data, in particular to a method for predicting crowd gathering trend based on an ARMA algorithm.
Background
When carrying out activities such as important activity prediction and public security, the aggregation of the concerned crowd can reflect the behavior of the concerned object, and meanwhile, the aggregation degree of the concerned crowd has a certain correlation with the occurrence of events in the object area. For this reason, it is necessary to be able to make an important study of the aggregation tendency of the crowd of interest.
Traditional research on the aggregation of people of interest based on artificial experience can achieve effects to a certain extent, but two obvious problems exist at the same time: first, for the ability to predict group events in advance, the concentration of the crowd of interest in traditional work often depends on the experience of researchers themselves, focusing on the management of crowd of interest in places where the events are frequent and people are intensive, but this approach cannot predict events effectively beyond experience; second, for setting the aggregation threshold, the aggregation threshold for the crowd of interest is manually set in the conventional work, and the criterion for judging the aggregation may change with the passage of time and the change of places, so the manual setting at this point cannot well divide the normal aggregation and the abnormal crowd event aggregation.
Disclosure of Invention
In order to find a more effective implementation scheme for crowd gathering trend prediction, the invention provides a method for predicting crowd gathering trend based on an ARMA algorithm, which comprises the following steps:
acquiring current day activity information of a preset place on a day to be predicted, and daily aggregate number values and daily activity information in p days before the day to be predicted, wherein p is a natural number;
obtaining a predicted aggregate value of the day to be predicted based on the current day activity information, the daily aggregate value in the p days and the daily activity information;
and acquiring a daily aggregate numerical value in r days before the day to be predicted, and confirming whether the predicted aggregate numerical value is abnormal or not based on the daily aggregate numerical value in r days and the predicted aggregate numerical value, if so, displaying that the predicted aggregate numerical value is abnormal, wherein r is a natural number not more than p.
Preferably, the identifying the predicted aggregate number value of the day to be predicted based on the day activity information and the daily aggregate number value and the daily activity information in the p days includes the steps of:
determining the activity level N of the day to be predicted according to a preset grading rule based on the activity information of the day t ；
Based on the activity level N t And training the daily aggregate person value and the daily activity information in the p days to obtain the following prediction function Y t ：
Wherein t is the day to be predicted, p is the number of days the day to be predicted t is pushed forward, and the p and q have the same numerical value;
y in (3) t-1 For the number of people gathered the day before the day t to be predicted,/->Is an aggregate weight;
as error term, e t-1 To predict the firstError value predicted on the previous day of t-day, < >>Is the error weight;
β t N t as an influence factor of the activity information of the day, beta t N is the activity weight t Is an activity level;
based on the prediction function Y t And obtaining the predicted aggregate value of the day to be predicted.
Preferably, the activity level is an integer from 0 to 4, and the activity level N of the day to be predicted is determined according to a preset classification rule based on the activity information of the day t The method comprises the following steps:
searching the number of activity participants of the same activity in a preset database based on the activity information of the same day, and if the number of activity participants cannot be found, determining the activity level N of the day to be predicted t Is 0;
if the activity participant number is found, reading the activity participant number nearest to the day to be predicted, and confirming the activity level N based on a preset grading rule t Wherein, the preset grading rule is as follows:
when the number of the activity participants is less than 500, the activity level N of the day to be predicted t 1 is shown in the specification;
when the number of the activity participants is not less than 500 people but less than 2000 people, the activity level N of the day to be predicted t Is 2;
when the number of the activity participants is not less than 2000 people but less than 5000 people, the activity level N of the day to be predicted t 3;
when the number of the activity participants is not less than 5000, the activity level N of the day to be predicted t 4.
Preferably, the aggregate weight and the error weight range is an integer from 0 to 10, based on the activity level N t And training the daily aggregate person value and the daily activity information in the p days to obtain a prediction function Y t The method comprises the following steps:
converting the daily aggregate numerical value and the daily activity information in the p days into matrixes, respectively assigning the aggregate weight and the initial RANDOM weight of the error weight to integers of 0-10 by using a RANDOM function, calculating a predicted value by forward propagation, and calculating a Loss value Loss function between the predicted value and a true value by using a least square method:
wherein y is i As a result of the pre-evaluation value,is a true value;
iterative obtaining the final weight parameter theta of the aggregation weight or the error weight according to a preset termination rule and a gradient descent method 1 ：
Wherein θ 0 For an initial random weight of the aggregate weight or the error weight,for the inverse derivative of each drop of the Loss function, +.>Is the learning rate.
Preferably, the preset termination rule includes one of the following conditions:
the average Loss value of the Loss function is not more than 1;
the number of iterations reaches fifty thousand times and the last ten times the reduction in the average loss value is less than 0.001.
Preferably, the saidThe value of (2) is 0.1.
Preferably, said determining whether said predicted aggregate value is an outlier based on said daily aggregate value and said predicted aggregate value over said r days comprises the steps of:
randomly selecting an aggregator value from the daily aggregator values and the predicted aggregator values over the r days;
according to the selected aggregate numerical value, carrying out data segmentation on the daily aggregate numerical value and the predicted aggregate numerical value in the r days in an isolated forest, placing records with the aggregate numerical value smaller than the selected aggregate numerical value on a left daughter, and placing records with the aggregate numerical value larger than or equal to the selected aggregate numerical value on a right child;
the recursive construction of the left and right daughter is stopped until one of the following conditions is met: the number of the incoming gatherer is only one record or a plurality of identical records; the height of the tree reaches a preset height;
and judging whether only one predicted aggregate number value is returned and whether the predicted aggregate number value is larger than a preset early warning value, if so, confirming that the predicted aggregate number value is an abnormal point.
Preferably, the preset height is one quarter of the total data volume of the daily aggregate value and the predicted aggregate value over the r days.
Preferably, the preset early warning value is the sum of the average value of the daily aggregate values in the r days and three times of the standard deviation of the daily aggregate values in the r days.
Preferably, r is 30 and p is 365.
Compared with the prior art, the method for predicting the crowd gathering trend based on the ARMA algorithm has the following beneficial effects:
according to the method for predicting the crowd gathering trend based on the ARMA algorithm, long-time multidimensional crowd gathering data of interest places are adopted, the ARMA algorithm is utilized to extract potential association between events which cannot be judged manually and time, comprehensive regression calculation is carried out by adding the event dimension, the gathering number of the next time point is predicted, and the defect of experience judgment of the crowd gathering degree of interest in the traditional work is overcome. Meanwhile, according to the crowd gathering condition of attention of historical data, the abnormal points in the crowd gathering condition are extracted by using an isolated forest algorithm to serve as abnormal early warning, a relatively fixed threshold value is not needed, and normal gathering and abnormal crowd event gathering can be effectively distinguished.
Additional aspects and advantages of the invention will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention.
Drawings
The foregoing and/or additional aspects and advantages of the invention will become apparent and readily appreciated from the following description of the embodiments, taken in conjunction with the accompanying drawings, in which:
FIG. 1 is a flow chart of a method for predicting crowd gathering tendency based on ARMA algorithm according to the embodiment of the invention;
Detailed Description
Embodiments of the present invention are described in detail below, examples of which are illustrated in the accompanying drawings, wherein like or similar reference numerals refer to like or similar elements or elements having like or similar functions throughout. The embodiments described below by referring to the drawings are illustrative only and are not to be construed as limiting the invention.
Referring to fig. 1, the method for predicting crowd gathering tendency based on ARMA algorithm of the present invention includes the following steps:
step S101: and acquiring current day activity information of a day to be predicted of a preset place, and daily aggregate number values and daily activity information in p days before the day to be predicted, wherein p is a natural number.
In some embodiments, the daily activity information or daily activity information is an exhibition, a literature show, a large meeting, a hot event, or the like. The preset places are places such as internet bars, hotels, airports, railway stations, crowd monitoring points, road bayonets and the like.
Notably, the crowd of interest and the preset place gathering data can be associated by taking the identity card number as an association field; holiday large-scale activity data is associated with historical venue aggregate data, the two data relate to time and space association, and the time can be directly associated, but no field similar to a preset venue code is associated in space (the venue-related dimension comprises longitude and latitude, venue name and venue address), so that comprehensive association is needed, and the method comprises the following specific steps of:
if the difference between longitude and latitude is within + -0.00001 degree (about 1.113 m in actual length), dividing the two places into the same place;
if some longitudes and latitudes do not find the matched places, extracting keywords from place names and place addresses by using jieba word segmentation and tf-idf algorithm, performing fuzzy matching on the keywords, calculating matching scores of the place names and the place addresses, adding score normalization, and dividing the highest score into the same places.
Meanwhile, when acquiring data, if there is a missing value, it is preferable to correlate the missing value with the existing data, for example, if a certain piece of data has a place code missing place name, the place name can be correlated with the place code. If the method is not relevant, the KNN nearest neighbor method is used for filling, and the embodiment of the invention is not described in detail.
Step S103: and obtaining the predicted aggregate value of the day to be predicted based on the daily aggregate value and the daily activity information in the current day and the p days.
Specifically, the determination of the predicted aggregate number value for the day to be predicted based on the day activity information and the daily aggregate number value and the daily activity information within p days includes the steps of:
determining an activity level N of a day to be predicted according to a preset grading rule based on the activity information of the day t ；
Based on activity level N t And daily aggregate person values and daily activity information within p days are trained to obtain the following predictive function Y t ：
Wherein t is the day to be predicted, p is the number of days the day to be predicted t is pushed forward, and the p and q have the same numerical value;
y in (3) t-1 For the number of people gathered the day before the day t to be predicted,/->Is an aggregate weight;
as error term, e t-1 For predicting the error value predicted on the day preceding the t-th day,/for the prediction of the error value on the day preceding the t-th day>Is the error weight;
β t N t as an influence factor of the activity information of the day, beta t N is the activity weight t Is an activity level;
based on predictive function Y t And obtaining the predicted aggregate value of the day to be predicted.
Preferably, the activity level is an integer from 0 to 4, and the activity level N of the day to be predicted is determined according to a preset classification rule based on the activity information of the day t The method comprises the following steps:
searching the number of activity participants of the same activity in a preset database based on the activity information of the same day, and if the number of activity participants cannot be found, predicting the activity level N of the day t Is 0;
if the number of the activity participants closest to the day to be predicted is found, reading the number of the activity participants closest to the day to be predicted, and confirming the activity level N based on a preset grading rule t Wherein, the preset grading rule is as follows:
when the number of the activity participants is less than 500, the activity level N of the day to be predicted is t 1 is shown in the specification;
when the number of the activity participants is not less than 500 people but less than 2000 people, the activity level N of the day to be predicted is t Is 2;
when the number of the activity participants is not less than 2000 people but less than 5000 people, the activity level N of the day to be predicted t 3;
when the number of the activity participants is not less than 5000, the activity level N of the day to be predicted t 4.
In some embodiments, the aggregate weight and the error weight range from an integer of 0 to 10 based on the activity level N t And training the daily aggregate person value and daily activity information in p days to obtain a prediction function Y t The method comprises the following steps:
the daily aggregate value and daily activity information in p days are converted into matrixes, the RANDOM function is used for respectively assigning the initial RANDOM weights of the aggregate weight and the error weight to integers of 0-10, after the forward propagation is used for calculating the predicted value, the least square method is used for calculating a Loss value Loss function between the predicted value and the true value:
wherein y is i As a result of the pre-evaluation value,is a true value;
iterative obtaining final weight parameter theta of the aggregate weight or the error weight according to a preset termination rule and a gradient descent method 1 ：
Wherein θ 0 To aggregate the initial random weights of the weights or error weights,for the inverse derivative of each drop of the Loss function, +.>Is the learning rate.
Preferably, the preset termination rule includes one of the following conditions: the average Loss value of the Loss function is not more than 1; the number of iterations reaches fifty thousand times and the last ten times the reduction in the average loss value is less than 0.001.
Preferably, the method comprises the steps of,the value of (2) is 0.1.
Step S105: and acquiring a daily aggregate numerical value in r days before the day to be predicted, and confirming whether the predicted aggregate numerical value is abnormal or not based on the daily aggregate numerical value in r days and the predicted aggregate numerical value, if so, displaying that the predicted aggregate numerical value is abnormal, wherein r is a natural number not more than p.
Specifically, determining whether the predicted aggregate value is an outlier based on the daily aggregate value and the predicted aggregate value within r days includes the steps of:
randomly selecting an aggregator value from the daily aggregator values and the predicted aggregator values within r days;
according to the selected aggregate numerical value, carrying out data segmentation on the isolated forest of the daily aggregate numerical value and the predicted aggregate numerical value in r days, placing records with the aggregate numerical value smaller than the selected aggregate numerical value on a left parapet, and placing records with the aggregate numerical value larger than or equal to the selected aggregate numerical value on a right child;
the recursive construction of the left and right daughter is stopped until one of the following conditions is met: the number of the incoming gatherer is only one record or a plurality of identical records; the height of the tree reaches a preset height;
and judging whether only one predicted aggregate number value is returned and whether the predicted aggregate number value is larger than a preset early warning value, if so, confirming that the predicted aggregate number value is an abnormal point.
Preferably, the preset height is one quarter of the total data volume of the daily aggregate value and the predicted aggregate value within r days. The preset early warning value is the sum of the average value of the daily aggregate values in r days and three times of the standard deviation of the daily aggregate values in r days.
In some embodiments, r is 30 and p is 365.
Compared with the prior art, the method for predicting the crowd gathering trend based on the ARMA algorithm has the following beneficial effects:
according to the method for predicting the crowd gathering trend based on the ARMA algorithm, long-time multidimensional crowd gathering data of interest places are adopted, the ARMA algorithm is utilized to extract potential association between events which cannot be judged manually and time, comprehensive regression calculation is carried out by adding the event dimension, the gathering number of the next time point is predicted, and the defect of experience judgment of the crowd gathering degree of interest in the traditional work is overcome. Meanwhile, according to the crowd location gathering condition of attention of historical data, the embodiment of the invention uses an isolated forest algorithm to extract the abnormal points in the crowd location gathering condition as abnormal early warning, does not need a relatively fixed threshold value, and can effectively distinguish normal gathering and abnormal crowd event gathering.
The foregoing is only a partial embodiment of the present invention, and it should be noted that it will be apparent to those skilled in the art that modifications and adaptations can be made without departing from the principles of the present invention, and such modifications and adaptations are intended to be comprehended within the scope of the present invention.
Claims (5)
1. The method for predicting the crowd gathering trend based on the ARMA algorithm is characterized by comprising the following steps of:
acquiring current day activity information of a preset place on a day to be predicted, and daily aggregate number values and daily activity information in p days before the day to be predicted, wherein p is a natural number;
obtaining a predicted aggregate value of the day to be predicted based on the current day activity information, the daily aggregate value in the p days and the daily activity information;
acquiring a daily aggregate number value in r days before the day to be predicted, and determining whether the predicted aggregate number value is abnormal or not based on the daily aggregate number value in r days and the predicted aggregate number value, if so, displaying that the predicted aggregate number value is abnormal, wherein r is a natural number not greater than p;
the step of confirming the predicted aggregate number value of the day to be predicted based on the day activity information and the daily aggregate number value and the daily activity information in the p days comprises the following steps:
determining the activity level N of the day to be predicted according to a preset grading rule based on the activity information of the day t ；
Based on the activity level N t And training the daily aggregate person value and the daily activity information in the p days to obtain the following prediction function Y t ：
Wherein t is the day to be predicted, p is the number of days the day to be predicted t is pushed forward, and the p and q have the same numerical value;
y in (3) t-1 For the number of people gathered the day before the day t to be predicted,/->Is an aggregate weight;
as error term, e t-1 For predicting the error value predicted on the day preceding the t-th day,/for the prediction of the error value on the day preceding the t-th day>Is the error weight;
β t N t as an influence factor of the activity information of the day, beta t N is the activity weight t Is an activity level;
based on the prediction function Y t Obtaining the predicted aggregate value of the day to be predicted;
the activity level is an integer of 0-4, and the activity level N of the day to be predicted is determined according to a preset grading rule based on the current day activity information t The method comprises the following steps:
searching the number of activity participants of the same activity in a preset database based on the activity information of the same day, and if the number of activity participants cannot be found, determining the activity level N of the day to be predicted t Is 0;
if the activity participant number is found, reading the activity participant number nearest to the day to be predicted, and confirming the activity level N based on a preset grading rule t Wherein, the preset grading rule is as follows:
when the number of the activity participants is less than 500, the activity level N of the day to be predicted t 1 is shown in the specification;
when the number of the activity participants is not less than 500 people but less than 2000 people, the activity level N of the day to be predicted t Is 2;
when the number of the activity participants is not less than 2000 people but less than 5000 people, the activity level N of the day to be predicted t 3;
when the number of the activity participants is not less than 5000, the activity level N of the day to be predicted t 4;
the value range of the aggregation weight and the error weight is an integer of 0-10, and the method is based on the activity level N t And training the daily aggregate person value and the daily activity information in the p days to obtain a prediction function Y t The method comprises the following steps:
converting the daily aggregate numerical value and the daily activity information in the p days into matrixes, respectively assigning the aggregate weight and the initial RANDOM weight of the error weight to integers of 0-10 by using a RANDOM function, calculating a predicted value by forward propagation, and calculating a Loss value Loss function between the predicted value and a true value by using a least square method:
wherein y is i As a result of the pre-evaluation value,is trueReal values;
iterative obtaining the final weight parameter theta of the aggregation weight or the error weight according to a preset termination rule and a gradient descent method 1 ：
Wherein θ 0 For an initial random weight of the aggregate weight or the error weight,for the inverse derivative of each drop of the Loss function, +.>Is the learning rate;
the preset termination rule includes one of the following conditions:
the average Loss value of the Loss function is not more than 1;
the iteration times reach fifty thousand times, and the reduction of the average loss value of the last ten times is less than 0.001;
the saidThe value of (2) is 0.1.
2. The method for predicting crowd-gathering tendency based on the ARMA algorithm as recited in claim 1, wherein said determining whether said predicted crowd-gathering value is an outlier based on said daily crowd-gathering value and said predicted crowd-gathering value over said r-days comprises the steps of:
randomly selecting an aggregator value from the daily aggregator values and the predicted aggregator values over the r days;
according to the selected aggregate numerical value, carrying out data segmentation on the daily aggregate numerical value and the predicted aggregate numerical value in the r days in an isolated forest, placing records with the aggregate numerical value smaller than the selected aggregate numerical value on a left daughter, and placing records with the aggregate numerical value larger than or equal to the selected aggregate numerical value on a right child;
the recursive construction of the left and right daughter is stopped until one of the following conditions is met: the number of the incoming gatherer is only one record or a plurality of identical records; the height of the tree reaches a preset height;
and judging whether only one predicted aggregate number value is returned and whether the predicted aggregate number value is larger than a preset early warning value, if so, confirming that the predicted aggregate number value is an abnormal point.
3. The method of predicting population gathering trend based on the ARMA algorithm as set forth in claim 2, wherein the preset height is one-fourth of the total data volume of the daily aggregate person number and the predicted aggregate person number over the r days.
4. The method for predicting crowd gathering trend based on the ARMA algorithm as recited in claim 2, wherein the preset pre-alarm value is a sum of an average value of the daily gathering people values in the r days and three times a standard deviation of the daily gathering people values in the r days.
5. The method for predicting crowd gathering tendency based on the ARMA algorithm as recited in claim 1 wherein r is 30 and p is 365.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010588491.1A CN111768031B (en) | 2020-06-24 | 2020-06-24 | Method for predicting crowd gathering trend based on ARMA algorithm |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010588491.1A CN111768031B (en) | 2020-06-24 | 2020-06-24 | Method for predicting crowd gathering trend based on ARMA algorithm |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111768031A CN111768031A (en) | 2020-10-13 |
| CN111768031B true CN111768031B (en) | 2023-09-19 |
Family
ID=72722390
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010588491.1A Active CN111768031B (en) | 2020-06-24 | 2020-06-24 | Method for predicting crowd gathering trend based on ARMA algorithm |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111768031B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113993153B (en) * | 2021-10-27 | 2023-07-04 | 中国联合网络通信集团有限公司 | Crowd gathering prediction method, device, equipment and computer readable storage medium |
Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| KR20000056762A (en) * | 1999-02-25 | 2000-09-15 | 이종수 | Prediction method for city water request of water service system |
| JP2004280731A (en) * | 2003-03-19 | 2004-10-07 | Nri & Ncc Co Ltd | Traffic congestion prediction system, traffic congestion prediction method, and compensation data system |
| CN107038492A (en) * | 2016-02-04 | 2017-08-11 | 滴滴（中国）科技有限公司 | Daily Order volume Forecasting Methodology and device based on Arma models |
| CN107798409A (en) * | 2016-08-30 | 2018-03-13 | 中兴智能交通股份有限公司 | A kind of crowd massing Forecasting Methodology based on time series models |
| CN108427989A (en) * | 2018-06-12 | 2018-08-21 | 中国人民解放军国防科技大学 | Deep space-time prediction neural network training method for radar echo extrapolation |
| CN108536652A (en) * | 2018-03-15 | 2018-09-14 | 浙江大学 | A kind of short-term vehicle usage amount prediction technique based on arma modeling |
| CN109002904A (en) * | 2018-06-21 | 2018-12-14 | 中南大学 | A kind of medical amount prediction technique of the hospital outpatient based on Prophet-ARMA |
| CN109492788A (en) * | 2017-09-13 | 2019-03-19 | 中移（杭州）信息技术有限公司 | Prediction flow of the people and the method and relevant device for establishing flow of the people prediction model |
| CN109697207A (en) * | 2018-12-25 | 2019-04-30 | 苏州思必驰信息科技有限公司 | The abnormality monitoring method and system of time series data |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20080033991A1 (en) * | 2006-08-03 | 2008-02-07 | Jayanta Basak | Prediction of future performance of a dbms |
-
2020
- 2020-06-24 CN CN202010588491.1A patent/CN111768031B/en active Active
Patent Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| KR20000056762A (en) * | 1999-02-25 | 2000-09-15 | 이종수 | Prediction method for city water request of water service system |
| JP2004280731A (en) * | 2003-03-19 | 2004-10-07 | Nri & Ncc Co Ltd | Traffic congestion prediction system, traffic congestion prediction method, and compensation data system |
| CN107038492A (en) * | 2016-02-04 | 2017-08-11 | 滴滴（中国）科技有限公司 | Daily Order volume Forecasting Methodology and device based on Arma models |
| CN107798409A (en) * | 2016-08-30 | 2018-03-13 | 中兴智能交通股份有限公司 | A kind of crowd massing Forecasting Methodology based on time series models |
| CN109492788A (en) * | 2017-09-13 | 2019-03-19 | 中移（杭州）信息技术有限公司 | Prediction flow of the people and the method and relevant device for establishing flow of the people prediction model |
| CN108536652A (en) * | 2018-03-15 | 2018-09-14 | 浙江大学 | A kind of short-term vehicle usage amount prediction technique based on arma modeling |
| CN108427989A (en) * | 2018-06-12 | 2018-08-21 | 中国人民解放军国防科技大学 | Deep space-time prediction neural network training method for radar echo extrapolation |
| CN109002904A (en) * | 2018-06-21 | 2018-12-14 | 中南大学 | A kind of medical amount prediction technique of the hospital outpatient based on Prophet-ARMA |
| CN109697207A (en) * | 2018-12-25 | 2019-04-30 | 苏州思必驰信息科技有限公司 | The abnormality monitoring method and system of time series data |
Non-Patent Citations (5)
| Title |
|---|
| ARMA模型在景区客流量预测中的应用;冯玉香;浙江统计(第10期);8-10 * |
| Enhancing Crowd Monitoring System Functionality through Data Fusion: Estimating Flow Rate from Wi-Fi Traces and Automated Counting System Data;Dorine C. Duives 等;sensors;第20卷(第21期);1-25 * |
| 基于ARMA和K-means聚类的用电量数据异常识别;梁捷;梁广明;;湖北电力(04);65-70 * |
| 基于航空大数据的机场客流量时空分布预测;罗甘;;电子技术与软件工程(18);155-156 * |
| 开源信息在突发事件应急管理中的应用;曾大军 等;科技导报;第26卷(第16期);27-33 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN111768031A (en) | 2020-10-13 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Ali et al. | A data-driven approach for multi-scale GIS-based building energy modeling for analysis, planning and support decision making | |
| US8065257B2 (en) | System and method for correlating past activities, determining hidden relationships and predicting future activities | |
| CN107977673B (en) | Economic activity population identification method based on big data | |
| WO2021004344A1 (en) | Data analysis-based risk identification method and related device | |
| CN111008337B (en) | Deep attention rumor identification method and device based on ternary characteristics | |
| TW201426578A (en) | Generation method and device and risk assessment method and device for anonymous dataset | |
| CN111950937A (en) | Key personnel risk assessment method based on fusion space-time trajectory | |
| WO2023168781A1 (en) | Soil cadmium risk prediction method based on spatial-temporal interaction relationship | |
| CN108595582B (en) | Social signal-based identification method for disastrous weather hot events | |
| CN110414715B (en) | Community detection-based passenger flow volume early warning method | |
| CN111126437B (en) | Abnormal group detection method based on weighted dynamic network representation learning | |
| Geetha et al. | Time-series modelling and forecasting: Modelling of rainfall prediction using ARIMA model | |
| CN113378990A (en) | Traffic data anomaly detection method based on deep learning | |
| CN112766119A (en) | Method for accurately identifying strangers and constructing community security based on multi-dimensional face analysis | |
| CN111768031B (en) | Method for predicting crowd gathering trend based on ARMA algorithm | |
| Syeed et al. | Flood prediction using machine learning models | |
| CN111831706A (en) | Mining method and device for association rules among applications and storage medium | |
| CN116226103A (en) | Method for detecting government data quality based on FPGrow algorithm | |
| Wang et al. | An algorithm for mining of association rules for the information communication network alarms based on swarm intelligence | |
| CN113254580A (en) | Special group searching method and system | |
| Ahani et al. | A feature weighting and selection method for improving the homogeneity of regions in regionalization of watersheds | |
| Ramakers et al. | Escaping the family tradition: A multi-generation study of occupational status and criminal behaviour | |
| CN115809280A (en) | Group house renting identification and iteration identification method | |
| CN113408867B (en) | Urban burglary crime risk assessment method based on mobile phone user and POI data | |
| KR102387284B1 (en) | Apparatus and method for forecasting heatwave Impact considering severity of health impacts and socio-economic vulnerability |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |