CN111008721A - Self-adaptive prediction method and system based on time series prediction model - Google Patents
Self-adaptive prediction method and system based on time series prediction model Download PDFInfo
- Publication number
- CN111008721A CN111008721A CN201811167942.3A CN201811167942A CN111008721A CN 111008721 A CN111008721 A CN 111008721A CN 201811167942 A CN201811167942 A CN 201811167942A CN 111008721 A CN111008721 A CN 111008721A
- Authority
- CN
- China
- Prior art keywords
- time
- adaptive
- time series
- prediction
- prediction model
- 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.)
- Pending
Links
Images
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
- 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
- G06Q10/10—Office automation; Time management
- G06Q10/105—Human resources
Landscapes
- Business, Economics & Management (AREA)
- Engineering & Computer Science (AREA)
- Human Resources & Organizations (AREA)
- Strategic Management (AREA)
- Economics (AREA)
- Entrepreneurship & Innovation (AREA)
- Operations Research (AREA)
- Marketing (AREA)
- Quality & Reliability (AREA)
- Tourism & Hospitality (AREA)
- Physics & Mathematics (AREA)
- General Business, Economics & Management (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Development Economics (AREA)
- Game Theory and Decision Science (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention belongs to the technical field of information, and particularly relates to a self-adaptive prediction method and a self-adaptive prediction system based on a time series prediction model. The method comprises the steps of obtaining a data time sequence of a prediction object within a certain period time; analyzing and processing the acquired data time sequence; decomposing the analyzed and processed data time sequence to obtain factor variables; carrying out algorithm improvement on the time series prediction model according to the factor variables; and designing a self-adaptive algorithm of a time window according to the improved result, predicting and obtaining a predicted result. By adopting the self-adaptive prediction method and the self-adaptive prediction system based on the time sequence prediction model, the time sequence prediction model can fully utilize information of seasonal periodicity of the time sequence with the percentage of departure and ensure a time window to realize self-adaptive selection of an optimal value in real time, and the prediction performance is high in stability and accuracy.
Description
Technical Field
The invention belongs to the technical field of information, and particularly relates to a self-adaptive prediction method and a self-adaptive prediction system based on a time series prediction model.
Background
The job leaving rate is an important index of the human operation condition of the company, and prediction of the job leaving rate can provide a very important reference basis for dealing with talent loss and human management and control of the company. Considering the outlier data as a time series, predicting the data using a time series prediction model (ARIMA) is a very common prediction algorithm.
However, in the process of actually using the algorithm, firstly, due to the existence of obvious seasonal periodicity of the percentage of departure, the applicability and the prediction stability of the algorithm are not high enough, and secondly, the time window period of the prediction signal also has a certain influence on the ARIMA model, so that the predicted results of different time window periods have differences. In conclusion, the ARIMA model cannot fully utilize seasonal periodicity information of the time series of the departure rate and the modeling time window cannot adaptively select the optimal value, so that the ARIMA model has the defects of low stability, low accuracy and the like in the prediction performance of the departure rate.
Disclosure of Invention
The invention provides a self-adaptive prediction method and a self-adaptive prediction system based on a time sequence prediction model, which aim to solve the problems that an ARIMA model cannot fully utilize seasonal periodicity information of a time sequence of an escape rate and a modeling time window cannot self-adaptively select an optimal value, so that the ARIMA model has the defects of low stability, low accuracy and the like in the prediction performance of the escape rate.
In order to solve the technical problem, the embodiment of the invention adopts the following technical scheme:
in one aspect, an embodiment of the present invention provides an adaptive prediction method based on a time series prediction model, where the method includes:
acquiring a data time sequence of a prediction object within a certain period of time; analyzing and processing the acquired data time sequence; decomposing the analyzed and processed data time sequence to obtain factor variables; carrying out algorithm improvement on the time series prediction model according to the factor variables; designing a self-adaptive algorithm of a time window according to the improved result, predicting and obtaining a predicted result;
further, before the analysis processing, preprocessing the data time sequence;
further, the preprocessing comprises one or more of dynamic data missing value filling and processing of outliers;
further, performing stability test and pure randomness test on the preprocessed data time sequence;
further, the analysis processing eliminates long-term trends and periodic changes by applying an artificial intelligence algorithm, and then solves the seasonal index according to a certain time period;
further, decomposing the analyzed and processed data time sequence through a decomposition model to obtain factor variables;
further, the factor variables include long-term trend factor variables, periodic variation factor variables and irregular factor variables;
further, the time series prediction model comprises an ARIMA model;
further, when the self-adaptive algorithm of the time window is designed, the current time window is determined to be the optimal time window in real time;
further, the prediction object is employee percentage of leave, and the data time sequence within the certain period of time is the percentage of leave time sequence of each month;
further, the predicted target is a product sales amount, and the data time series within the certain period time is a product sales amount time series of each quarter.
In a second aspect, an embodiment of the present invention further provides an adaptive prediction system based on a time series prediction model, including: a data source module, an analysis processing module, a decomposition module, a prediction module and a result output module,
the data source module is used for acquiring a data time sequence of the prediction object within a certain period time;
the analysis processing module is used for analyzing and processing the acquired data time sequence;
the decomposition module is used for decomposing the analyzed and processed data time sequence so as to obtain factor variables of the data time sequence;
the prediction module is used for improving the time series prediction model through factor variables, designing a self-adaptive algorithm of a time window and predicting;
and the result output module is used for outputting the prediction result.
The system further comprises a preprocessing module, wherein the preprocessing module is used for preprocessing the data time sequence obtained by the data source module;
and further, the system also comprises a verification module which is used for performing stability verification and pure randomness verification on the data time series before analysis processing.
The invention provides a self-adaptive prediction method and a self-adaptive prediction system based on a time sequence prediction model, which have the following beneficial effects: acquiring a data time sequence of a prediction object within a certain period of time; analyzing and processing the acquired data time sequence; decomposing the analyzed and processed data time sequence to obtain factor variables; carrying out algorithm improvement on the time series prediction model according to the factor variables; and designing a self-adaptive algorithm of a time window according to the improved result, predicting and obtaining a predicted result. By adopting the self-adaptive prediction method and the self-adaptive prediction system based on the time sequence prediction model, the time sequence prediction model can fully utilize information of seasonal periodicity of the time sequence with the percentage of departure and ensure a time window to realize self-adaptive selection of an optimal value in real time, and the prediction performance is high in stability and accuracy.
Drawings
FIG. 1 is a flow chart of an adaptive prediction method based on a time series prediction model according to an embodiment of the present invention;
FIG. 2 is a flowchart of an adaptive prediction method based on a time series prediction model according to an embodiment of the present invention;
FIG. 3 is a diagram of an adaptive prediction system based on a time series prediction model according to an embodiment of the present invention.
Detailed Description
The technical solutions in the present invention will be described clearly and completely with reference to the accompanying drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only some embodiments of the present invention, not all embodiments.
Detailed description of the preferred embodiment
The embodiment of the invention discloses a self-adaptive prediction method based on a time series prediction model, and a flow chart of the method is shown in figure 1 and comprises the following steps:
s1: acquiring a time sequence of the employee departure rate within each month;
the employee job leaving data is regarded as a data time sequence, and a job leaving rate time sequence of the employee job leaving rate in each month is obtained, wherein the job leaving rate time sequence shows seasonal periodicity in time;
s2: preprocessing the time sequence of the job leaving rate;
before analysis and processing, preprocessing the acquired time series of the off-duty rate, wherein the preprocessing comprises one or more of dynamic data missing value filling and processing of an outlier, and the integrity and the reliability of signals of the time series of the off-duty rate are ensured;
s3: stability test and pure randomness test;
performing stationarity test and pure randomness test on the preprocessed time sequence of the percentage of leaving, so as to ensure that the time sequence of the percentage of leaving has a usable time sequence prediction model for prediction;
s4: analyzing and processing the acquired time sequence of the job leaving rate;
analyzing and processing the acquired time sequence of the leaving percentage, eliminating long-term trend and periodic variation by using an artificial intelligence algorithm, and then solving a seasonal index according to a certain time period;
s5: decomposing the analyzed and processed time series of the job leaving rate to obtain factor variables;
decomposing the analyzed and processed time series of the job leaving rate through a decomposition model to obtain factor variables; the factor variables comprise long-term trend factor variables, periodic variation factor variables and irregular factor variables;
s6: carrying out algorithm improvement on the time series prediction model according to the factor variables;
performing algorithm improvement on the time series prediction model according to the long-term trend factor variable, the periodic variation factor variable, the irregular factor variable and other factor variables obtained from S5; the time sequence prediction model is an ARIMA model;
s7: designing a self-adaptive algorithm of a time window according to the improved result, predicting and obtaining a predicted result;
designing a time window self-adaptive algorithm according to the improved result, and determining the current time sequence time window as the optimal time window in real time, thereby realizing the on-line self-adaptability of the ARIMA model algorithm and improving the availability and universality of the ARIMA model algorithm; the predicted outcome includes employee attendance within a month in the future.
Detailed description of the invention
The embodiment of the invention discloses a self-adaptive prediction method based on a time series prediction model, and a flow chart of the method is shown in figure 2 and comprises the following steps:
s1: obtaining a product sales time series within each quarter of product sales;
taking the product sales data as a data time sequence, and acquiring the product sales time sequence in each quarter of the product sales, wherein the product sales time sequence shows seasonal periodicity in time;
s2: preprocessing a product sales time sequence;
before analysis processing, preprocessing the obtained product sales time sequence, wherein the preprocessing comprises one or more of dynamic data missing value filling and outlier processing, and the integrity and reliability of a product sales time sequence signal are ensured;
s3: stability test and pure randomness test;
performing stationarity test and pure randomness test on the preprocessed product sales time sequence to ensure that the product sales time sequence has a usable time sequence prediction model for prediction;
s4: analyzing and processing the obtained product sales time sequence;
analyzing and processing the obtained product sales time sequence, eliminating long-term trend and periodic variation by using an artificial intelligence algorithm, and then solving a seasonal index according to a certain time period;
s5: decomposing the product sales time sequence after analysis and treatment to obtain factor variables;
decomposing the product sales time series after analysis processing through a decomposition model, and obtaining factor variables through the decomposition model; the factor variables comprise long-term trend factor variables, periodic variation factor variables and irregular factor variables;
s6: carrying out algorithm improvement on the time series prediction model according to the factor variables;
performing algorithm improvement on the time series prediction model according to the long-term trend factor variable, the periodic variation factor variable, the irregular factor variable and other factor variables obtained from S5; the time sequence prediction model is an ARIMA model;
s7: designing a self-adaptive algorithm of a time window according to the improved result, predicting and obtaining a predicted result;
designing a time window self-adaptive algorithm according to the improved result, and determining the current time sequence time window as the optimal time window in real time, thereby realizing the on-line self-adaptability of the ARIMA model algorithm and improving the availability and universality of the ARIMA model algorithm; the forecasted results include product sales within a certain quarter in the future.
The embodiment of the invention discloses a self-adaptive prediction system based on a time series prediction model, the schematic diagram of the system is shown in figure 3, and the system comprises: the system comprises a data source module 1, a preprocessing module 2, a verification module 3, an analysis processing module 4, a decomposition module 5, a prediction module 6 and a result output module 7;
the data source module 1 is used for acquiring a data time sequence of a prediction object within a certain period time; the time series of data exhibits seasonal periodicity in time; if the prediction object is the employee attendance rate, acquiring an employee attendance rate time sequence of the employee attendance rate within a certain period of time; if the prediction object is the product sales volume, acquiring a product sales volume time sequence within a certain period time of the product sales volume;
the preprocessing module 2 is used for preprocessing the data time sequence obtained by the data source module 1; preprocessing comprises one or more of dynamic data missing value filling and outlier processing so as to ensure the integrity and reliability of data time series signals;
the verification module 3 is used for performing stationarity test and pure randomness test on the data time sequence; performing stationarity test and pure randomness test on the data time sequence processed by the preprocessing module 2 to ensure that the data time sequence has a usable prediction model for prediction;
the analysis processing module 4 is used for analyzing and processing the data time series; eliminating long-term trend and periodic variation by using an artificial intelligence algorithm, and then solving a seasonal index according to a certain time period;
the decomposition module 5 is used for decomposing the analyzed data time sequence and then obtaining factor variables of the data time sequence by applying a decomposition model; the factor variables comprise long-term trend factor variables, seasonal variation factor variables, periodic variation factor variables and irregular factor variables;
the prediction module 6 is used for carrying out algorithm improvement on the time series prediction model according to the factor variables, designing a self-adaptive algorithm of a time window and predicting;
performing algorithm improvement on the time series prediction model according to the long-term trend factor variable, the periodic variation factor variable, the irregular factor variable and other factor variables obtained by the decomposition module 5; the time sequence prediction model is an ARIMA model; designing a time window self-adaptive algorithm according to the improved result, and determining the current time sequence time window as the optimal time window in real time, thereby realizing the on-line self-adaptability of the ARIMA model algorithm and improving the availability and universality of the ARIMA model algorithm;
a result output module 7, configured to output a prediction result; the prediction result is a prediction target in a certain period of time in the future, such as: employee turnover in a month in the future, product sales in a quarter in the future.
The above description is only a preferred embodiment of the present invention, but the design concept of the present invention is not limited thereto, and any insubstantial modifications made by using the design concept should fall within the scope of infringing on the protection scope of the present invention.
Claims (14)
1. An adaptive prediction method based on a time series prediction model is characterized in that:
acquiring a data time sequence of a prediction object within a certain period of time;
analyzing and processing the acquired data time sequence;
decomposing the analyzed and processed data time sequence to obtain factor variables;
carrying out algorithm improvement on the time series prediction model according to the factor variables;
and designing a self-adaptive algorithm of a time window according to the improved result, predicting and obtaining a predicted result.
2. The adaptive prediction method based on time series prediction model according to claim 1, characterized in that: and preprocessing the data time sequence before the analysis processing.
3. The adaptive prediction method based on time series prediction model according to claim 2, characterized in that: the preprocessing comprises one or more of dynamic data missing value filling and processing of outliers.
4. The adaptive prediction method based on time series prediction model according to claim 2, characterized in that: and performing stationarity test and pure randomness test on the preprocessed data time sequence.
5. The adaptive prediction method based on time series prediction model according to claim 1, characterized in that: the analysis processing eliminates long-term trend and periodic variation by applying an artificial intelligence algorithm, and then solves the seasonal index according to a certain time period.
6. The adaptive prediction method based on time series prediction model according to claim 1, characterized in that: and decomposing the analyzed and processed data time sequence through a decomposition model to obtain factor variables.
7. The adaptive prediction method based on time series prediction model according to claim 1, characterized in that: the factor variables include long-term trend factor variables, periodic variation factor variables, and irregular factor variables.
8. The adaptive prediction method based on time series prediction model according to claim 1, characterized in that: the time series prediction model comprises an ARIMA model.
9. The adaptive prediction method based on time series prediction model according to claim 1, characterized in that: and when the self-adaptive algorithm of the time window is designed, determining the current time window as the optimal time window in real time.
10. The adaptive prediction method based on time series prediction model according to claim 1, characterized in that: the prediction object is the employee percentage of leaving, and the data time sequence within a certain period of time is the percentage of leaving time sequence of each month.
11. The adaptive prediction method based on time series prediction model according to claim 1, characterized in that: the forecast target is product sales, and the data time series in the certain period time is the product sales time series of each quarter.
12. An adaptive prediction system based on a time series prediction model is characterized in that: comprises a data source module, an analysis processing module, a decomposition module, a prediction module and a result output module,
the data source module is used for acquiring a data time sequence of the prediction object within a certain period time;
the analysis processing module is used for analyzing and processing the acquired data time sequence;
the decomposition module is used for decomposing the analyzed and processed data time sequence so as to obtain factor variables of the data time sequence;
the prediction module is used for improving the time series prediction model through factor variables, designing a self-adaptive algorithm of a time window and predicting;
and the result output module is used for outputting the prediction result.
13. The adaptive prediction system based on time series prediction model according to claim 12, wherein: the system also comprises a preprocessing module, wherein the preprocessing module is used for preprocessing the data time sequence obtained by the data source module.
14. The adaptive prediction system based on time series prediction model according to claim 12, wherein: the system also comprises a verification module used for carrying out stationarity test and pure randomness test on the data time sequence before analysis and processing.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201811167942.3A CN111008721A (en) | 2018-10-08 | 2018-10-08 | Self-adaptive prediction method and system based on time series prediction model |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201811167942.3A CN111008721A (en) | 2018-10-08 | 2018-10-08 | Self-adaptive prediction method and system based on time series prediction model |
Publications (1)
Publication Number | Publication Date |
---|---|
CN111008721A true CN111008721A (en) | 2020-04-14 |
Family
ID=70111132
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201811167942.3A Pending CN111008721A (en) | 2018-10-08 | 2018-10-08 | Self-adaptive prediction method and system based on time series prediction model |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN111008721A (en) |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN112633594A (en) * | 2020-12-30 | 2021-04-09 | 北京高思博乐教育科技股份有限公司 | Automatic prediction method, device and system for multi-target time sequence |
-
2018
- 2018-10-08 CN CN201811167942.3A patent/CN111008721A/en active Pending
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN112633594A (en) * | 2020-12-30 | 2021-04-09 | 北京高思博乐教育科技股份有限公司 | Automatic prediction method, device and system for multi-target time sequence |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
Ford et al. | Smart grid energy fraud detection using artificial neural networks | |
Kemerer et al. | The impact of design and code reviews on software quality: An empirical study based on psp data | |
CN110019401B (en) | Method, device, equipment and storage medium for predicting part quantity | |
US11190535B2 (en) | Methods and systems for inferring behavior and vulnerabilities from process models | |
CN110390425A (en) | Prediction technique and device | |
CN111639783A (en) | Line loss prediction method and system based on LSTM neural network | |
Harper et al. | A hybrid modelling approach using forecasting and real-time simulation to prevent emergency department overcrowding | |
CN116663746A (en) | Power load prediction method and device, computer equipment and storage medium | |
CN111932292A (en) | Cigarette product sales prediction method and device based on deep neural network | |
CN115936895A (en) | Risk assessment method, device and equipment based on artificial intelligence and storage medium | |
CN111080472A (en) | Load prediction and analysis method for power system | |
Gajda et al. | Modeling of water usage by means of ARFIMA–GARCH processes | |
CN111008721A (en) | Self-adaptive prediction method and system based on time series prediction model | |
CN115545362B (en) | New energy medium-term power prediction method combining artificial intelligence and time sequence decomposition | |
CN112700050A (en) | Method and system for predicting ultra-short-term 1 st point power of photovoltaic power station | |
CN111738507A (en) | Bank clearing position fund payment amount prediction method, device, equipment and medium | |
CN112766535B (en) | Building load prediction method and system considering load curve characteristics | |
CN113962741B (en) | Coal sales data prediction method, equipment and medium | |
CN111105148B (en) | Off-job probability evaluation method, apparatus and computer readable storage medium | |
CN114463086A (en) | E-commerce information security method combining big data and readable storage medium | |
Friederich et al. | A Framework for Validating Data-Driven Discrete-Event Simulation Models of Cyber-Physical Production Systems | |
CN109754175B (en) | Computational model for compressed prediction of transaction time limit of administrative examination and approval items and application thereof | |
Crone et al. | Feature selection of autoregressive neural network inputs for trend time series forecasting | |
Kosztyán et al. | Computer aided diagnostic methods to forecast condition-based maintenance tasks | |
CN111815458A (en) | Dynamic investment portfolio configuration method based on fine-grained quantitative marking and integration method |
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 | ||
RJ01 | Rejection of invention patent application after publication | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20200414 |