CN110580639A - Store operation prediction method based on chain business state - Google Patents
Store operation prediction method based on chain business state Download PDFInfo
- Publication number
- CN110580639A CN110580639A CN201910730327.7A CN201910730327A CN110580639A CN 110580639 A CN110580639 A CN 110580639A CN 201910730327 A CN201910730327 A CN 201910730327A CN 110580639 A CN110580639 A CN 110580639A
- Authority
- CN
- China
- Prior art keywords
- prediction
- model
- data
- samples
- class
- 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.)
- Withdrawn
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
- G06Q30/00—Commerce
- G06Q30/02—Marketing; Price estimation or determination; Fundraising
- G06Q30/0201—Market modelling; Market analysis; Collecting market data
- G06Q30/0202—Market predictions or forecasting for commercial activities
Landscapes
- Business, Economics & Management (AREA)
- Strategic Management (AREA)
- Engineering & Computer Science (AREA)
- Accounting & Taxation (AREA)
- Development Economics (AREA)
- Finance (AREA)
- Entrepreneurship & Innovation (AREA)
- Game Theory and Decision Science (AREA)
- Data Mining & Analysis (AREA)
- Economics (AREA)
- Marketing (AREA)
- Physics & Mathematics (AREA)
- General Business, Economics & Management (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention provides a chain-state-based store operation prediction method, which comprises the steps of generating a dash board scheme through data preliminary processing and visualization, comparing data prediction effects under different models through multiple prediction models, evaluating the prediction models through secondary indexes and tertiary indexes to obtain a model most suitable for characteristic data, and finally comparing model prediction data with reality data through data backtracking comparison to continuously adjust parameters of the model so as to continuously update and evolve the model.
Description
Technical Field
The invention particularly relates to a store operation prediction method based on chain business states.
Background
the operation prediction analysis refers to economic benefits and development trends of future operation activities, and the scientific prediction special analysis is proposed in advance, so that the necessity and the possible process of business development are predicted by using the mastered scientific knowledge and the practical experience of managers on the basis of the past historical data and the currently available information. The predictive analysis serves the decision making and is a precondition for the scientization of the decision making. The prior art on the market at present lacks visual show to data, can not form the technique closed loop simultaneously, lacks the technique backtracking.
disclosure of Invention
The invention aims to provide a chain-business-based store operation prediction method, which integrates a BI data analysis strategy to enable data analysis to be more visual and data comparison to be more obvious.
the invention provides the following technical scheme:
a store operation prediction method based on chain business states comprises the following steps:
S1, performing primary processing on the original historical data and the peripheral external environment data;
s2, constructing a BI system from a plurality of angles by using a mathebase BI tool, and performing visualization processing on the historical data to generate a DashBoard;
s3, carrying out ring comparison and comparing on the data under different dimensions to obtain obvious operation influence factors, and making an analysis result into a dash board to finish the preprocessing of the data;
s4, selecting a plurality of prediction models, adding the preprocessed data into the prediction models for prediction deduction, and obtaining the prediction effect of the data under each prediction model;
S5, evaluating the prediction model according to the secondary index and the tertiary index according to the prediction effect, selecting the prediction model suitable for the historical data and taking the prediction model as an operation condition prediction model;
And S6, comparing the prediction data of the prediction model with the actual data after new data are obtained, and modifying the model parameters according to the difference between the prediction data and the actual data to realize data backtracking.
2. the chain business-based store operation prediction method according to claim 1, wherein the preliminary processing of step S1 includes data regularization and data null value removal.
3. The chain business model-based store operation prediction method according to claim 1, wherein the plurality of angles in step S2 include a store closing analysis plan, a company summary, a store continuation analysis plan, a new store opening plan and a prediction new analysis.
Preferably, the influencing factors of step S3 include internal factors including store revenue, store area and store level, and external factors including passenger flow, average consumption and average person GDP.
Preferably, the prediction model in step S4 includes a random forest model, a gradient lifting tree model, a common least square method model, a ridge regression model, a lasso regression model, a decision tree model, a support vector machine model, a random gradient descent model, a K-nearest neighbor algorithm model, a random tree model, a neural network model, and a lasso path model.
preferably, the secondary indicators in step S5 include accuracy, precision, sensitivity and specificity, the accuracy is the percentage of the total samples that are predicted to be correct, the precision is the probability that all samples that are predicted to be positive are actually positive, the sensitivity is the probability that all samples that are predicted to be positive, and the specificity is the probability that all samples that are actually negative are predicted to be positive.
preferably, the calculation formula of the accuracy is as follows: (TP + TN)/(TP + TN + FP + FN), the calculation formula of the accuracy rate is as follows: TP/(TP + FP), the calculation formula of the sensitivity is as follows: TP/(TP + FN), the calculation formula of the specificity is as follows: TN/(TN + FP), where TP is the number of samples for which the true class of samples is the positive class and the prediction result of the prediction model is the number of samples for the positive class, TN is the number of samples for which the true class of samples is the negative class and the prediction result of the prediction model is the number of samples for the negative class, FP is the number of samples for which the true class of samples is the negative class and the prediction result of the prediction model is the number of samples for the positive class, FN is the number of samples for which the true class of samples is the negative class and the prediction result of the prediction model is the number of samples for the positive class.
Preferably, the three-level index in step S4 includes F1Score and AUC/ROC characteristic curve, and the value range of the F1Score is 0-1.
Preferably, the calculation formula of the F1Score is as follows: 2PR/(P + R), wherein P is precision and R is sensitivity.
Preferably, the abscissa of the ROC characteristic curve is: 1-specificity, and the ordinate of the ROC characteristic curve is sensitivity.
The invention has the beneficial effects that: according to the invention, through data preliminary processing, visualization is realized on data by adopting a mathebase BI tool, a dash board scheme is generated, the distribution condition of the data under various dimensions is effectively preliminarily shown for a client, and the problem that the data can only be manually retrieved in the past is solved; through the application of multiple prediction models, data prediction effects under different kinds of prediction models are compared, and a secondary index and a tertiary index are adopted for evaluation, so that a model most suitable for characteristic data is obtained, and the problem of overlarge model prediction deviation is solved; through data backtracking comparison, model prediction data is compared with reality data, so that parameters of the model are continuously adjusted, and the model is continuously updated and evolved.
Drawings
the accompanying drawings, which are included to provide a further understanding of the invention and are incorporated in and constitute a part of this specification, illustrate embodiments of the invention and together with the description serve to explain the principles of the invention and not to limit the invention. In the drawings:
FIG. 1 is an architectural diagram of the present invention;
FIG. 2 is a sample graph of data visualization of the present invention;
FIG. 3 is a graph of the comparison of geo-market GDP to per-person consumption according to the present invention;
FIG. 4 is a graph of the effect of the model prediction of the present invention;
Detailed Description
as shown in fig. 1, a method for predicting store operation based on chain state comprises the following steps:
S1, performing primary processing on the original historical data and the peripheral external environment data, wherein the primary processing comprises data regularization and data null value removal;
S2, constructing a BI system from the perspectives of store closing analysis plan, company summary, store continuation analysis plan, new store opening plan, prediction new increase analysis and the like by using a mathebase BI tool, and performing visualization processing on historical data to generate Dashboards, wherein each Dash Board comprises a large number of analysis charts, and the data comparison visualization sample diagram is shown in FIG. 2;
s3, carrying out ring comparison and identity comparison on the data under different dimensions to obtain obvious operation influence factors, wherein the influence factors comprise internal factors and external factors, the internal factors comprise store revenues, store areas and store levels, the external factors comprise passenger flow, average consumption and average people Group (GDP), the analysis result is manufactured into a dash board, and the data preprocessing is completed;
s4, selecting a plurality of prediction models, adding the preprocessed data into the prediction models for prediction deduction to obtain the prediction effect of the data under each prediction model, wherein the selected prediction models comprise Random Forest models, Gradient boost Trees, Ordinary Least Squares, Ridge Regression models, Lasso Regression models, XGboost decision Trees, Support Vector machines, chacGraded decision Trees, KNN (K neighbor algorithm models), Extra Random Trees, Neural networks and Lasso paths;
s5, evaluating the prediction model according to the second-level index and the third-level index according to the prediction effect, selecting the prediction model suitable for the historical data and taking the prediction model as an operation condition prediction model;
specifically, the secondary indicators include accuracy, precision, sensitivity, and specificity, the accuracy being the percentage of the total samples that are predicted to be correct, the precision being the probability of actually being positive among all samples that are predicted to be positive, the sensitivity being the probability of predicting to be positive among all samples that are predicted to be positive, and the specificity being the probability of being negative among samples that are actually negative,
Wherein, the calculation formula of the accuracy is as follows: (TP + TN)/(TP + TN + FP + FN), the calculation formula of the accuracy rate is as follows: TP/(TP + FP), the sensitivity is calculated according to the formula: TP/(TP + FN), the calculation formula of specificity is: TN/(TN + FP), where TP is the number of samples for which the true category of the sample is a positive category and the prediction result of the prediction model is the number of samples for which the positive category is a negative category, TN is the number of samples for which the true category of the sample is a negative category and the prediction result of the prediction model is the number of samples for which the true category of the sample is a negative category, FP is the number of samples for which the true category of the sample is a negative category and the prediction result of the prediction model is the number of samples for which the true category of the sample is a negative category;
Specifically, the three-level index includes F1Score and an AUC/ROC characteristic curve, the value range of the F1Score is 0-1, and specifically, the calculation formula of the F1Score is as follows: 2PR/(P + R), wherein P is the precision and R is the sensitivity, and for F1Score, the closer the value is to 1, the better the model output is; the abscissa of the ROC characteristic curve is: 1-specificity, and the ordinate of the ROC characteristic curve is sensitivity, then the calculation formulas of the abscissa and the ordinate are respectively FP/(FP + TN) and TP/(TP + FN), where TP is the sample true category is the positive category, and the prediction result of the prediction model is the sample number of the positive category, TN is the sample true category is the negative category, and the prediction result of the prediction model is the sample number of the negative category, FP is the sample true category is the negative category, and the prediction result of the prediction model is the sample number of the positive category, FN is the sample true category is the negative category, and the prediction result of the prediction model is the sample number of the positive category;
And S6, comparing the prediction data of the prediction model with the actual data after the new data are obtained, and modifying the model parameters according to the difference between the prediction data and the actual data to realize data backtracking.
The first embodiment is as follows:
(1) Historical data of the business of the customer provider and peripheral external environment data;
(2) Primary processing client provided data (data regularization, data minus null values, etc.);
(3) Performing visualization processing on data from the aspects of store closing analysis plan, company summary, store continuing analysis plan, new store opening plan and prediction and new increase analysis to generate Dash boards, wherein each Dash Board comprises a large number of analysis charts;
(4) Dividing store operation influence factors into external factors and internal factors, wherein the external influence factors comprise peripheral passenger flow, peripheral per-capita consumption, city GDP (global distribution platform) of the city, and the internal influence factors comprise store area, decoration condition, store level and the like;
(5) the data were ring-to-ring compared in different dimensions, the effect of the comparison is shown in figure 3,
(6) The analysis result is manufactured into a dash board through preliminary comparison and analysis, so that the distribution condition of data under various dimensions is effectively preliminarily shown for a client, and the problem that the data can only be manually retrieved in the past is solved;
(7) adding the preprocessed data into a plurality of prediction models to perform prediction deduction, obtaining prediction effects of the data under each prediction model, and evaluating the model effects, wherein the model prediction effects are shown in FIG. 4, so that the problem of overlarge model prediction deviation is solved;
(8) After the actual operation data of the current year is obtained, the actual operation data is compared with the model prediction operation data to find out that the model is insufficient in prediction, data backtracking is carried out, and model parameters are adjusted to enable the model to be continuously updated and evolved.
although the present invention has been described in detail with reference to the foregoing embodiments, it will be apparent to those skilled in the art that changes may be made in the embodiments and/or equivalents thereof without departing from the spirit and scope of the invention. Any modification, equivalent replacement, or improvement made within the spirit and principle of the present invention should be included in the protection scope of the present invention.
Claims (10)
1. A store operation prediction method based on chain business states comprises the following steps:
s1, performing primary processing on the original historical data and the peripheral external environment data;
s2, constructing a BI system from a plurality of angles by using a mathebase BI tool, and performing visualization processing on the historical data to generate a DashBoard;
s3, carrying out ring comparison and comparing on the data under different dimensions to obtain obvious operation influence factors, and making an analysis result into a dash board to finish the preprocessing of the data;
S4, selecting a plurality of prediction models, adding the preprocessed data into the prediction models for prediction deduction, and obtaining the prediction effect of the data under each prediction model;
S5, evaluating the prediction model according to the secondary index and the tertiary index according to the prediction effect, selecting the prediction model suitable for the historical data and taking the prediction model as an operation condition prediction model;
and S6, comparing the prediction data of the prediction model with the actual data after new data are obtained, and modifying the model parameters according to the difference between the prediction data and the actual data to realize data backtracking.
2. The chain business-based store operation prediction method according to claim 1, wherein the preliminary processing of step S1 includes data regularization and data null value removal.
3. The chain business model-based store operation prediction method according to claim 1, wherein the plurality of angles in step S2 include a store closing analysis plan, a company summary, a store continuation analysis plan, a new store opening plan and a prediction new analysis.
4. the chain business state-based store operation prediction method according to claim 1, wherein the influencing factors of step S3 include internal factors and external factors, the internal factors include store revenue, store area and store level, and the external factors include passenger flow, average consumption and average people GDP.
5. The chain-business-state-based store operation prediction method according to claim 1, wherein the prediction models in step S4 include a random forest model, a gradient lifting tree model, a common least square model, a ridge regression model, a lasso regression model, a decision tree model, a support vector machine model, a random gradient descent model, a K-nearest neighbor algorithm model, a random tree model, a neural network model, and a lasso path model.
6. the method of claim 1, wherein the secondary measures in step S5 include accuracy, precision, sensitivity and specificity, the accuracy is a percentage of a total sample that is predicted to be correct, the precision is a probability that all samples that are predicted to be positive are actually positive, the sensitivity is a probability that all samples that are predicted to be positive, and the specificity is a probability that all samples that are actually negative are predicted to be positive.
7. the chain business state-based store operation prediction method according to claim 6, wherein the accuracy is calculated by the formula: (TP + TN)/(TP + TN + FP + FN), the calculation formula of the accuracy rate is as follows: TP/(TP + FP), the calculation formula of the sensitivity is as follows: TP/(TP + FN), the calculation formula of the specificity is as follows: TN/(TN + FP), where TP is the number of samples for which the true class of samples is the positive class and the prediction result of the prediction model is the number of samples for the positive class, TN is the number of samples for which the true class of samples is the negative class and the prediction result of the prediction model is the number of samples for the negative class, FP is the number of samples for which the true class of samples is the negative class and the prediction result of the prediction model is the number of samples for the positive class, FN is the number of samples for which the true class of samples is the negative class and the prediction result of the prediction model is the number of samples for the positive class.
8. the method of claim 1, wherein the three-level index of step S4 includes F1Score and AUC/ROC characteristic curve, and the F1Score has a value ranging from 0 to 1.
9. the chain business state-based store operation prediction method according to claim 8, wherein the calculation formula of F1Score is: 2PR/(P + R), wherein P is precision and R is sensitivity.
10. the chain business state-based store operation prediction method according to claim 8, wherein the abscissa of the ROC characteristic curve is: 1-specificity, and the ordinate of the ROC characteristic curve is sensitivity.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910730327.7A CN110580639A (en) | 2019-08-08 | 2019-08-08 | Store operation prediction method based on chain business state |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910730327.7A CN110580639A (en) | 2019-08-08 | 2019-08-08 | Store operation prediction method based on chain business state |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN110580639A true CN110580639A (en) | 2019-12-17 |
Family
ID=68810864
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910730327.7A Withdrawn CN110580639A (en) | 2019-08-08 | 2019-08-08 | Store operation prediction method based on chain business state |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110580639A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111198499A (en) * | 2019-12-25 | 2020-05-26 | 南京南瑞水利水电科技有限公司 | Synchronous algorithm real-time evaluation method, system and storage medium |
| CN113538021A (en) * | 2020-04-09 | 2021-10-22 | 上海丙晟科技有限公司 | Machine learning algorithm for store continuity prediction of shopping mall |
| TWI793412B (en) * | 2020-03-31 | 2023-02-21 | 廣達電腦股份有限公司 | Consumption prediction system and consumption prediction method |
-
2019
- 2019-08-08 CN CN201910730327.7A patent/CN110580639A/en not_active Withdrawn
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111198499A (en) * | 2019-12-25 | 2020-05-26 | 南京南瑞水利水电科技有限公司 | Synchronous algorithm real-time evaluation method, system and storage medium |
| TWI793412B (en) * | 2020-03-31 | 2023-02-21 | 廣達電腦股份有限公司 | Consumption prediction system and consumption prediction method |
| US11983726B2 (en) | 2020-03-31 | 2024-05-14 | Quanta Computer Inc. | Consumption prediction system and consumption prediction method |
| CN113538021A (en) * | 2020-04-09 | 2021-10-22 | 上海丙晟科技有限公司 | Machine learning algorithm for store continuity prediction of shopping mall |
| CN113538021B (en) * | 2020-04-09 | 2023-11-10 | 上海丙晟科技有限公司 | Machine learning method for store duration prediction |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107045788B (en) | Traffic road condition prediction method and device | |
| CN110580639A (en) | Store operation prediction method based on chain business state | |
| CN112052992B (en) | Deep learning-based construction project progress prediction system and method | |
| CN109685240A (en) | A kind of prediction of 95598 traffic work orders and unusual fluctuation method for early warning based on LSTM deep learning | |
| CN106254317A (en) | A kind of data security exception monitoring system | |
| CN107742168A (en) | A kind of workshop bottleneck Forecasting Methodology based on Internet of Things technology | |
| CN106530082A (en) | Stock predication method and stock predication system based on multi-machine learning | |
| CN113537807B (en) | Intelligent wind control method and equipment for enterprises | |
| CN116822769A (en) | Intelligent traffic route optimizing system based on artificial intelligence | |
| CN114118507A (en) | Risk assessment early warning method and device based on multi-dimensional information fusion | |
| CN116264388A (en) | Short-term load prediction method based on GRU-LightGBM model fusion and Bayesian optimization | |
| CN112257914B (en) | Aviation safety causal prediction method based on random forest | |
| CN115907611B (en) | Fitting inventory control method based on fitting market value | |
| CN116957331A (en) | Risk passenger flow range prediction method and device | |
| CN111898867A (en) | Airplane final assembly production line productivity prediction method based on deep neural network | |
| CN114548494B (en) | Visual cost data prediction intelligent analysis system | |
| CN115796936A (en) | Cigarette sales prediction method and system based on combined model and storage medium | |
| CN117787569B (en) | Intelligent auxiliary bid evaluation method and system | |
| Tianyang | A K-means group division and LSTM based method for hotel demand forecasting | |
| CN114548350A (en) | Power load prediction method based on goblet sea squirt group and BP neural network | |
| CN104217296A (en) | Listed company performance comprehensive evaluation method | |
| CN109697531A (en) | A kind of logistics park-hinterland Forecast of Logistics Demand method | |
| CN115660296B (en) | Automatic issuing method of non-compliance project proposal scheme based on machine learning | |
| CN106779245A (en) | Civil aviaton's needing forecasting method and device based on event | |
| Ruta et al. | Diversified gradient boosting ensembles for prediction of the cost of forwarding contracts |
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 | ||
| WW01 | Invention patent application withdrawn after publication | ||
| WW01 | Invention patent application withdrawn after publication |
Application publication date: 20191217 |