CN114118637A - Method and device for building prediction model of accessory demand and computer equipment - Google Patents
Method and device for building prediction model of accessory demand and computer equipment Download PDFInfo
- Publication number
- CN114118637A CN114118637A CN202210104179.XA CN202210104179A CN114118637A CN 114118637 A CN114118637 A CN 114118637A CN 202210104179 A CN202210104179 A CN 202210104179A CN 114118637 A CN114118637 A CN 114118637A
- Authority
- CN
- China
- Prior art keywords
- data
- prediction
- prediction model
- accessory
- demand
- 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
- 238000000034 method Methods 0.000 title claims abstract description 60
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 35
- 238000012549 training Methods 0.000 claims abstract description 18
- 238000010801 machine learning Methods 0.000 claims abstract description 15
- 230000002159 abnormal effect Effects 0.000 claims description 30
- 238000012545 processing Methods 0.000 claims description 13
- 238000004590 computer program Methods 0.000 claims description 12
- 238000003860 storage Methods 0.000 claims description 9
- 238000010276 construction Methods 0.000 claims description 8
- 230000009469 supplementation Effects 0.000 claims description 6
- 238000013528 artificial neural network Methods 0.000 claims description 4
- 238000012937 correction Methods 0.000 claims description 3
- 239000000126 substance Substances 0.000 claims description 3
- 230000001502 supplementing effect Effects 0.000 claims description 3
- 238000010586 diagram Methods 0.000 description 11
- 230000006870 function Effects 0.000 description 4
- 238000007726 management method Methods 0.000 description 4
- 238000002474 experimental method Methods 0.000 description 3
- 238000004519 manufacturing process Methods 0.000 description 2
- 238000006467 substitution reaction Methods 0.000 description 2
- 240000007594 Oryza sativa Species 0.000 description 1
- 235000007164 Oryza sativa Nutrition 0.000 description 1
- 238000012896 Statistical algorithm Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000007405 data analysis Methods 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 239000003337 fertilizer Substances 0.000 description 1
- 238000009472 formulation Methods 0.000 description 1
- 238000012417 linear regression Methods 0.000 description 1
- 238000012423 maintenance Methods 0.000 description 1
- 230000014759 maintenance of location Effects 0.000 description 1
- 239000000203 mixture Substances 0.000 description 1
- YHXISWVBGDMDLQ-UHFFFAOYSA-N moclobemide Chemical compound C1=CC(Cl)=CC=C1C(=O)NCCN1CCOCC1 YHXISWVBGDMDLQ-UHFFFAOYSA-N 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000007637 random forest analysis Methods 0.000 description 1
- 235000009566 rice Nutrition 0.000 description 1
- 238000005096 rolling process Methods 0.000 description 1
- 238000012706 support-vector machine Methods 0.000 description 1
- 230000001131 transforming effect Effects 0.000 description 1
- 238000011144 upstream manufacturing Methods 0.000 description 1
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
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- 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
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/04—Manufacturing
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
- Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
- Y02P90/30—Computing systems specially adapted for manufacturing
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Business, Economics & Management (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Strategic Management (AREA)
- General Health & Medical Sciences (AREA)
- Evolutionary Computation (AREA)
- Life Sciences & Earth Sciences (AREA)
- General Engineering & Computer Science (AREA)
- Human Resources & Organizations (AREA)
- Economics (AREA)
- Health & Medical Sciences (AREA)
- Artificial Intelligence (AREA)
- Software Systems (AREA)
- Computational Linguistics (AREA)
- Biophysics (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- General Business, Economics & Management (AREA)
- Biomedical Technology (AREA)
- Tourism & Hospitality (AREA)
- Marketing (AREA)
- Bioinformatics & Computational Biology (AREA)
- Game Theory and Decision Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Entrepreneurship & Innovation (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- Development Economics (AREA)
- Evolutionary Biology (AREA)
- Manufacturing & Machinery (AREA)
- Primary Health Care (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The application provides a method and a device for building a prediction model of accessory requirements and computer equipment. The method comprises the following steps: acquiring basic data corresponding to accessories containing different reference types; dividing basic data into a plurality of data sets through a time sliding window with preset length, wherein each data set comprises different characteristic factors, and each characteristic factor comprises historical demand data corresponding to the (n-i) th preset period to the nth preset period and covariate data corresponding to the (n +1) th preset period; taking all data sets as training sets, and performing regression prediction through various machine learning algorithms to obtain corresponding reference prediction models; and selecting a target prediction model with the highest prediction precision corresponding to each type of accessory from all the reference prediction models. According to the method, the characteristic factors are selected on the basis of the historical demand data and the covariate data of each accessory, the training set of the prediction model is divided according to the characteristic factors, the characteristic engineering covering the demand characteristics of the accessories is constructed, and the method is high in applicability and accuracy.
Description
Technical Field
The application relates to the technical field of computers, in particular to a method and a device for building a prediction model of accessory requirements and computer equipment.
Background
In a supply chain system of the manufacturing industry, demand forecasting of after-sales accessories is a key plan, which directly affects inventory planning and supply capacity of upstream and downstream enterprises such as accessory suppliers and distributors, and has a great influence on operation indexes such as fund occupation and spot-stock satisfaction rate. The sparsity of historical data of accessory sales increases the difficulty of accessory demand prediction.
Currently, in the field of demand forecasting of after-market accessories in manufacturing industry, the conventional solutions are as follows: (1) relying on manual experience: working personnel compile a demand prediction value of the accessory through historical sales experience by means of Excel and other tools; (2) with the aid of an informatization tool: the information management tools of the modeling enterprises with a certain scale or more have high popularization rate, for example, an ERP system is used, workers complete warehouse entry and exit management and data statistics of after-sales accessories by the aid of the system, and the quantity of the required accessories is predicted by the aid of a simple statistical algorithm.
However, the above scheme has the following disadvantages: (1) the method excessively depends on the experience and the capability of an accessory demand formulation personnel, so that the precision of demand prediction is insufficient, and the prediction method is not reusable; (2) the ERP system and other information management tools can complete the management of the parts in and out of the warehouse, but the adopted algorithm for demand prediction of the parts is simple, the prediction precision is low, and for different types of parts, the historical demand data has different characteristics, and the customized prediction algorithm cannot be adopted.
Disclosure of Invention
In order to solve the technical problem, the application provides a method, a device and a computer device for building a prediction model of accessory requirements, and the specific scheme is as follows:
in a first aspect, an embodiment of the present application provides a method for building a prediction model of an accessory demand, where the method for building a prediction model of an accessory demand includes:
acquiring basic data corresponding to accessories containing different reference types, wherein the basic data comprise historical demand data and covariate data which take a preset period as a unit;
dividing the basic data into a plurality of data sets through a time sliding window with preset length, wherein each data set comprises different characteristic factors, and each characteristic factor comprises the historical demand data corresponding to the (n-i) th preset period to the nth preset period and the covariate data corresponding to the (n +1) th preset period, wherein n and i are positive integers, and n > i;
taking all the data sets as training sets, and performing regression prediction through multiple machine learning algorithms to obtain reference prediction models corresponding to the machine learning algorithms;
and selecting a target prediction model with the highest prediction precision corresponding to each type of accessory from all the reference prediction models.
According to a specific embodiment disclosed in the present application, the step of obtaining basic data corresponding to accessories containing different reference types includes:
acquiring original data corresponding to accessories containing different reference types;
judging whether the covariate data in the original data is complete or not, wherein the covariate data comprises a plurality of sub-data with strong correlation;
and completely supplementing the covariate data through a data supplementation algorithm to obtain complete basic data corresponding to the original data, wherein the data supplementation algorithm comprises a Transformer neural network algorithm.
According to a specific embodiment disclosed in the present application, the sub-data includes a total storage amount, a total operation amount, and a total operation duration in units of the preset period.
According to a specific embodiment disclosed in the present application, the step of dividing the basic data into a plurality of data sets by a time sliding window of a preset length includes:
judging whether the basic data has abnormal data or not;
if the abnormal data exist in the basic data, performing deviation rectification on each abnormal data;
and dividing the basic data after the deviation rectification processing into a plurality of data sets through a time sliding window with a preset length.
According to a specific embodiment disclosed in the present application, the step of determining whether there is abnormal data in the basic data, where the abnormal data includes at least one of first data and second data, includes:
calculating a mean value and a standard deviation corresponding to the basic data, wherein the basic data comprises a plurality of reference data;
judging whether each reference data exceeds the sum of the mean value and the three standard deviations or whether each reference data is smaller than the difference of the mean value and the three standard deviations;
if any reference data exceeds the sum of the mean value and the three standard deviations, judging the corresponding reference data as the first data;
and if any reference data is smaller than the difference value between the mean value and the three standard deviations, judging the corresponding reference data as the second data.
According to a specific embodiment disclosed in the present application, the step of performing error correction processing on each abnormal data includes:
judging whether the abnormal data is the first data or not;
if the abnormal data is the first numberAccording to the first formula
Assigning the first data as deviation correcting data;
if the abnormal data is the second data, passing a second formula
Assigning the first data to be the deviation correcting data;
wherein the content of the first and second substances,
for the assigned deviation correcting data, the deviation correcting data is calculated,
in order to be able to process the first data,
in order to be able to process the second data,
and the standard deviation corresponding to the basic data.
According to a specific embodiment disclosed in the present application, after the step of selecting a target prediction model with the highest prediction accuracy corresponding to each type of the accessory from all the reference prediction models, the method for constructing a prediction model of accessory demand further includes:
receiving a target type corresponding to an accessory input by a user and a target period to be predicted;
selecting the target prediction model corresponding to the target type as a standard prediction model;
acquiring the historical demand data corresponding to (i +1) continuous preset periods before the target period and the predicted value of the covariate data corresponding to the target period as reference data;
and inputting the reference data into the standard prediction model for prediction to obtain target demand data corresponding to the accessory of the target type.
In a second aspect, an embodiment of the present application provides a device for building a prediction model of an accessory demand, where the device for building a prediction model of an accessory demand includes:
the device comprises an acquisition module, a processing module and a processing module, wherein the acquisition module is used for acquiring basic data corresponding to accessories containing different reference types, and the basic data comprises historical demand data and covariate data which take a preset period as a unit;
the dividing module is used for dividing the basic data into a plurality of data sets through a time sliding window with preset length, wherein each data set comprises different characteristic factors, and each characteristic factor comprises the historical demand data corresponding to the (n-i) th preset period to the nth preset period and the covariate data corresponding to the (n +1) th preset period, wherein n and i are positive integers, and n > i;
the prediction module is used for performing regression prediction by using all the data sets as training sets through various machine learning algorithms to obtain reference prediction models corresponding to the machine learning algorithms;
and the selection module is used for selecting a target prediction model with the highest prediction precision corresponding to each type of accessory from all the reference prediction models.
In a third aspect, an embodiment of the present application provides a computer device, where the computer device includes a processor and a memory, where the memory stores a computer program, and the computer program, when executed on the processor, implements the method for building a predictive model of an accessory demand described in any one of the embodiments of the first aspect.
In a fourth aspect, the present application provides a computer-readable storage medium, where a computer program is stored, and when the computer program is executed on a processor, the computer program implements the method for building a predictive model of an accessory demand described in any one of the embodiments of the first aspect.
Compared with the prior art, the method has the following beneficial effects:
the method includes the steps that basic data corresponding to accessories with different reference types are obtained, wherein the basic data comprise historical demand data and covariate data which take a preset period as a unit. And constructing a characteristic factor by combining the historical demand data and the covariate data, constructing a data set for model training based on the characteristic factor, and performing model training through various algorithms to obtain a model with the highest prediction precision as a target prediction model. The method and the device can construct the characteristic engineering which fully covers the requirement characteristics of the accessories, and the target prediction model obtained by the characteristic engineering has strong applicability and high accuracy.
Drawings
In order to more clearly illustrate the technical solution of the present invention, the drawings required to be used in the embodiments will be briefly described below, and it should be understood that the following drawings only illustrate some embodiments of the present invention, and therefore should not be considered as limiting the scope of the present invention. Like components are numbered similarly in the various figures.
Fig. 1 is a schematic flowchart of a method for building a prediction model of an accessory demand according to an embodiment of the present application;
fig. 2 is a schematic diagram of classification of raw data related to a method for building a prediction model of an accessory requirement according to an embodiment of the present application;
fig. 3 is one of schematic diagrams of time sliding windows involved in a method for building a prediction model of an accessory demand according to an embodiment of the present application;
fig. 4 is a second schematic diagram of a time sliding window involved in a method for building a prediction model of an accessory requirement according to an embodiment of the present application;
fig. 5 is a block diagram of a device for building a prediction model of an accessory demand according to an embodiment of the present application.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments.
The components of embodiments of the present invention generally described and illustrated in the figures herein may be arranged and designed in a wide variety of different configurations. Thus, the following detailed description of the embodiments of the present invention, presented in the figures, is not intended to limit the scope of the invention, as claimed, but is merely representative of selected embodiments of the invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments of the present invention without making any creative effort, shall fall within the protection scope of the present invention.
Hereinafter, the terms "including", "having", and their derivatives, which may be used in various embodiments of the present invention, are only intended to indicate specific features, numbers, steps, operations, elements, components, or combinations of the foregoing, and should not be construed as first excluding the existence of, or adding to, one or more other features, numbers, steps, operations, elements, components, or combinations of the foregoing.
Furthermore, the terms "first," "second," "third," and the like are used solely to distinguish one from another and are not to be construed as indicating or implying relative importance.
Unless otherwise defined, all terms (including technical and scientific terms) used herein have the same meaning as commonly understood by one of ordinary skill in the art to which various embodiments of the present invention belong. The terms (such as those defined in commonly used dictionaries) should be interpreted as having a meaning that is consistent with their contextual meaning in the relevant art and will not be interpreted in an idealized or overly formal sense unless expressly so defined herein in various embodiments of the present invention.
Some embodiments of the present application will be described in detail below with reference to the accompanying drawings. The embodiments described below and the features of the embodiments can be combined with each other without conflict.
Referring to fig. 1, fig. 1 is a schematic flowchart of a method for building a prediction model of an accessory demand according to an embodiment of the present application. As shown in fig. 1, the method for building a prediction model of accessory demand mainly includes:
step S101, acquiring basic data corresponding to accessories containing different reference types, wherein the basic data comprises historical demand data and covariate data which take a preset period as a unit.
The reference type is used for distinguishing different accessories and can be in various forms such as ID and name of the accessories. The basic data may be sales data or demand data or the like corresponding to all reference types of accessories in a certain period in the past. Covariates refer to an independent variable in the design of an experiment that is not manipulated by the experimenter but still affects the experimental results. For example, in an experiment, the influence of fertilizer on the yield of rice can be experimentally controlled, and the variables are not covariates. When the influence of training on income is researched, individual attribute variables such as gender and ethnicity, income level before training, economic and social development characteristics and the like are covariates, and the variables are not influenced by the training experiment. The covariate data is included in the construction process of the accessory prediction model, various factors influencing accessory requirements can be comprehensively considered, and the applicability and the accuracy of the accessory requirement model are improved.
The preset period can be time periods with different lengths such as day, week and month, and the user can customize the preset period according to actual use requirements and specific application scenes, and the preset period is not further limited here. And then dividing all the basic data into historical demand data and covariate data which take a preset period as a unit.
The method comprises the following steps of acquiring basic data corresponding to accessories containing different reference types, wherein the basic data comprises the following steps:
acquiring original data corresponding to accessories containing different reference types;
judging whether the covariate data in the original data is complete or not, wherein the covariate data comprises a plurality of sub-data with strong correlation;
and completely supplementing the covariate data through a data supplementation algorithm to obtain complete basic data corresponding to the original data, wherein the data supplementation algorithm comprises a Transformer neural network algorithm.
Referring to fig. 2, fig. 2 is a schematic diagram illustrating classification of raw data related to a method for building a prediction model of an accessory demand according to an embodiment of the present application. After acquiring raw data corresponding to accessories containing different reference types, there may be three different situations:
1. the time sequence data is complete, and the covariate data is complete;
2. the time sequence data is complete, and the covariate data is completely lost;
3. the time series data was complete and the covariate data was partially missing.
The time-series data refers to the historical demand data in the preset period unit as described above. If the covariate data is complete, the subsequent processing steps can be directly carried out; if the covariate data is partially lost, the lost covariate data can be complemented through data analysis and then subjected to subsequent processing. Specifically, data completion may be performed by the sub data having a strong correlation and the prediction algorithm. The data completion algorithm can adopt a Transformer neural network algorithm. After the data is supplemented, covariate data becomes complete, the subsequent model construction can be carried out by adopting the same prediction algorithm as the first kind of situation, and other different prediction algorithms can be selected for processing.
In the second category, the covariate data is completely missing and cannot be predicted from the historical demand data of the accessory. In this case, a differential Integrated Moving Average Autoregressive Model (ARIMA for short) may be used for the subsequent prediction step.
In specific implementation, the subdata constituting the covariate data includes the total maintenance number, the total operating number and the total operating duration of the whole machine, which take a preset period as a unit.
The reserve is the amount of a product that can be normally used in the market for a certain period of time in a durable consumer product or an industrial product. In short, the retention amount refers to how many such products are already on the market. In the case of automobiles, the holding amount refers to the number of automobiles owned by a certain area, and is generally vehicles registered in local related departments.
The whole machine is the whole machine corresponding to the various accessories. For example, the aforementioned various reference types of accessories may be various components that make up the a device, and the complete machine is the a device.
Step S102, dividing the basic data into a plurality of data sets through a time sliding window with preset length, wherein each data set comprises different characteristic factors, and each characteristic factor comprises the historical demand data corresponding to the (n-i) th preset period to the nth preset period and the covariate data corresponding to the (n +1) th preset period, wherein n and i are positive integers, and n > i.
The time sliding window is a window that can frame time series data according to a specified unit length, such as the historical demand data described above, so as to calculate statistical data in the frame. The slide block with the designated length slides on the scale, and data corresponding to the slide block can be fed back when the slide block slides one unit.
Referring to fig. 3, fig. 3 is one of schematic diagrams of time sliding windows involved in a method for building a prediction model of an accessory demand according to an embodiment of the present application. The time series data shown in the figure is composed of 12 data, and all of the 12 data can be sequentially divided into A, B, C, D and E5 data sets by a time sliding window with a length of 4 units.
If the performance of the constructed demand prediction model needs to be optimized, the best machine learning algorithm needs to be selected, and more information needs to be obtained from the original data as much as possible. The method is the corresponding practical meaning of the feature engineering, and aims to obtain better model training data. Feature engineering is a process of transforming raw data into features that can well describe the raw data. The characteristic factor is a very important step in the construction process of the characteristic engineering, namely, a data set with representative significance or statistical significance is selected from all original data.
Referring to fig. 4, fig. 4 is a second schematic diagram of a time sliding window involved in a method for building a prediction model of an accessory demand according to an embodiment of the present application. The original data can be scrolled by using a time sliding window with a preset length, and the original data is expanded into a plurality of data sets by combining covariate data. The preset length can be customized by a user according to actual use requirements and a specific application scene, and is not further limited here. The preset length of the time sliding window as shown in fig. 4 may correspond to raw data corresponding to 12 months, and the characteristic factor may be historical demand data corresponding to 12 months and covariate data corresponding to months required to be predicted. For example, knowing the sales volume, the total machine holding volume, the total machine startup volume and the total machine startup duration of 2018/1-2020/6 for 30 months, the window can be scrolled based on 12 months, and the input characteristic factors are: the sales volume of 2018/1-2018/12 is historical demand data, the total machine holding volume of 2019/1, the total machine operating quantity and the total machine operating time are 15 features, and 18 samples can be obtained by rolling 30 months of original data.
The method comprises the following steps of dividing the basic data into a plurality of data sets through a time sliding window with a preset length, wherein the steps comprise:
judging whether the basic data has abnormal data or not;
if the abnormal data exist in the basic data, performing deviation rectification on each abnormal data;
and dividing the basic data after the deviation rectification processing into a plurality of data sets through a time sliding window with a preset length.
The obtained covariate data in the original data corresponding to the accessories containing different reference types may be partially missing, and the missing covariate data may be complemented by a data complementing algorithm to obtain complete basic data. However, there is a possibility that abnormal data exists in the complete basic data. Whether abnormal data exists in the basic data or not can be judged through the following steps, and the abnormal data can be classified into first data and second data:
calculating a mean value and a standard deviation corresponding to the basic data, wherein the basic data comprises a plurality of reference data;
judging whether each reference data exceeds the sum of the mean value and the three standard deviations or whether each reference data is smaller than the difference of the mean value and the three standard deviations;
if any reference data exceeds the sum of the mean value and the three standard deviations, judging the corresponding reference data as the first data;
and if any reference data is smaller than the difference value between the mean value and the three standard deviations, judging the corresponding reference data as the second data.
In specific implementation, the limiting conditions involved in the above determining step, that is, "the sum of the mean value and the three standard deviations" and "the difference between the mean value and the three standard deviations", may be customized by the user according to actual use requirements and specific application scenarios, and are not further limited herein.
If the abnormal data exists in the basic data, the abnormal data can be processed according to different types of the abnormal data. Specifically, the step of performing error correction processing on each abnormal data includes:
judging whether the abnormal data is the first data or not;
if the abnormal data is the first data, passing a first formula
Assigning the first data as deviation correcting data;
if the abnormal data is the second data, passing a second formula
Assigning the first data to be the deviation correcting data;
wherein the content of the first and second substances,
for the assigned deviation correcting data, the deviation correcting data is calculated,
in order to be able to process the first data,
in order to be able to process the second data,
and the standard deviation corresponding to the basic data.
If the above-mentioned limiting conditions, that is, "the sum of the mean and the three standard deviations" and "the difference between the mean and the three standard deviations" are changed according to the actual usage requirement, the first formula and the second formula herein are also correspondingly adjusted, and are not described in detail herein.
And S103, taking all the data sets as training sets, and performing regression prediction through multiple machine learning algorithms to obtain reference prediction models corresponding to the machine learning algorithms.
In particular implementations, the machine learning algorithm may include at least one of linear regression, KNN regression, support vector machine regression, ridge regression, decision regression trees, random forests, and XGBOOST.
And step S104, selecting a target prediction model with the highest prediction precision corresponding to each type of accessory from all the reference prediction models.
After the step of determining the target prediction model, the target prediction model corresponding to the target type can be selected as the standard prediction model by receiving the target type corresponding to the accessory input by the user and the target period to be predicted. Acquiring historical demand data corresponding to (i +1) continuous preset periods before the target period and a predicted value of covariate data corresponding to the target period as reference data. And inputting the reference data into a standard prediction model for prediction to obtain target demand data corresponding to the accessory of the target type.
For example, if the demand data of a certain B reference type of accessory needs to be predicted in 2020/7 months, the target type at this time is the B reference type, and the target period to be predicted is 2020/7 months. The target prediction model with the highest prediction accuracy corresponding to the accessory of the reference type B can be selected from all the reference prediction models. The acquired reference data includes historical demand data corresponding to (i +1) consecutive preset periods before 2020/7, such as sales data of the B-reference type of accessories in 12 months before 2020/7, and predicted values corresponding to the total holding amount, the total operating number and the total operating duration in 2020/7 months.
In specific implementation, the total machine holding amount refers to the total machine number which is registered or counted, the use of the total machine is performed according to a preset construction plan or use plan, and the value generally does not have large fluctuation. Therefore, the predicted values corresponding to the whole machine holding amount, the whole machine start-up number and the whole machine start-up time length are conveniently obtained and are accurate. The reference data may be used as an input of the target prediction model, so as to output target demand data corresponding to 2020/7 months, that is, a predicted value.
The method includes the steps that basic data corresponding to accessories with different reference types are obtained, wherein the basic data comprise historical demand data and covariate data which take a preset period as a unit. And constructing a characteristic factor by combining the historical demand data and the covariate data, constructing a data set for model training based on the characteristic factor, and performing model training through various algorithms to obtain a model with the highest prediction precision as a target prediction model. The method and the device can construct the characteristic engineering which fully covers the requirement characteristics of the accessories, and the target prediction model obtained by the characteristic engineering has strong applicability and high accuracy.
Corresponding to the above method embodiment, referring to fig. 5, the present invention further provides a device 500 for building a prediction model of an accessory demand, where the device 500 for building a prediction model of an accessory demand includes:
an obtaining module 501, configured to obtain basic data corresponding to accessories including different reference types, where the basic data includes historical demand data and covariate data in units of preset periods;
a dividing module 502, configured to divide the basic data into multiple data sets through a time sliding window of a preset length, where each data set includes different feature factors, and each feature factor includes the historical demand data corresponding to (n-i) th to nth preset periods and the covariate data corresponding to (n +1) th preset period, where n and i are positive integers, and n > i;
the prediction module 503 is configured to perform regression prediction by using all the data sets as training sets through multiple machine learning algorithms to obtain reference prediction models corresponding to the machine learning algorithms;
a selecting module 504, configured to select, from all the reference prediction models, a target prediction model with the highest prediction accuracy corresponding to each type of accessory.
The application provides a prediction model construction device, computer equipment and a computer readable storage medium for accessory requirements, which can be used for acquiring basic data corresponding to accessories containing different reference types, wherein the basic data comprises historical requirement data and covariate data which take a preset period as a unit. And constructing a characteristic factor by combining the historical demand data and the covariate data, constructing a data set for model training based on the characteristic factor, and performing model training through various algorithms to obtain a model with the highest prediction precision as a target prediction model. The method and the device can construct the characteristic engineering which fully covers the requirement characteristics of the accessories, and the target prediction model obtained by the characteristic engineering has strong applicability and high accuracy.
In addition, a computer device is further provided, the computer device comprises a processor and a memory, the memory stores a computer program, and the computer program realizes the prediction model construction method of the accessory requirement when executed on the processor.
In addition, a computer-readable storage medium is provided, in which a computer program is stored, and the computer program realizes the above method for constructing the prediction model of the accessory demand when executed on a processor.
For specific implementation processes of the device for building a prediction model of an accessory demand, the computer device, and the computer-readable storage medium provided in the present application, reference may be made to the specific implementation processes of the method for building a prediction model of an accessory demand provided in the foregoing embodiments, and details are not described here any more.
In the embodiments provided in the present application, it should be understood that the disclosed apparatus and method can be implemented in other ways. The apparatus embodiments described above are merely illustrative and, for example, the flowchart and block diagrams in the figures illustrate the architecture, functionality, and operation of possible implementations of apparatus, methods and computer program products according to various embodiments of the present invention. In this regard, each block in the flowchart or block diagrams may represent a module, segment, or portion of code, which comprises one or more executable instructions for implementing the specified logical function(s). It should also be noted that, in alternative implementations, the functions noted in the block may occur out of the order noted in the figures. For example, two blocks shown in succession may, in fact, be executed substantially concurrently, or the blocks may sometimes be executed in the reverse order, depending upon the functionality involved. It will also be noted that each block of the block diagrams and/or flowchart illustration, and combinations of blocks in the block diagrams and/or flowchart illustration, can be implemented by special purpose hardware-based systems which perform the specified functions or acts, or combinations of special purpose hardware and computer instructions.
In addition, each functional module or unit in each embodiment of the present invention may be integrated together to form an independent part, or each module may exist separately, or two or more modules may be integrated to form an independent part.
The functions, if implemented in the form of software functional modules and sold or used as a stand-alone product, may be stored in a computer readable storage medium. Based on such understanding, the technical solution of the present invention or a part of the technical solution that contributes to the prior art in essence can be embodied in the form of a software product, which is stored in a storage medium and includes instructions for causing a computer device (which may be a smart phone, a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the method according to the embodiments of the present invention. And the aforementioned storage medium includes: a U-disk, a removable hard disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), a magnetic disk or an optical disk, and other various media capable of storing program codes.
The above description is only for the specific embodiments of the present invention, but the scope of the present invention is not limited thereto, and any person skilled in the art can easily conceive of the changes or substitutions within the technical scope of the present invention, and all the changes or substitutions should be covered within the scope of the present invention.
Claims (10)
1. A method for building a prediction model of an accessory demand is characterized by comprising the following steps:
acquiring basic data corresponding to accessories containing different reference types, wherein the basic data comprise historical demand data and covariate data which take a preset period as a unit;
dividing the basic data into a plurality of data sets through a time sliding window with preset length, wherein each data set comprises different characteristic factors, and each characteristic factor comprises the historical demand data corresponding to the (n-i) th preset period to the nth preset period and the covariate data corresponding to the (n +1) th preset period, wherein n and i are positive integers, and n > i;
taking all the data sets as training sets, and performing regression prediction through multiple machine learning algorithms to obtain reference prediction models corresponding to the machine learning algorithms;
and selecting a target prediction model with the highest prediction precision corresponding to each type of accessory from all the reference prediction models.
2. The method for building a prediction model of accessory demand according to claim 1, wherein the step of obtaining basic data corresponding to accessories containing different reference types comprises:
acquiring original data corresponding to accessories containing different reference types;
judging whether the covariate data in the original data is complete or not, wherein the covariate data comprises a plurality of sub-data with strong correlation;
and completely supplementing the covariate data through a data supplementation algorithm to obtain complete basic data corresponding to the original data, wherein the data supplementation algorithm comprises a Transformer neural network algorithm.
3. The method for building a prediction model of an accessory demand according to claim 2, wherein the sub-data includes a total holding quantity, a total start-up quantity, and a total start-up duration in units of the preset period.
4. The method for building a prediction model of accessory demand according to claim 1, wherein the step of dividing the basic data into a plurality of data sets by a time sliding window of a preset length comprises:
judging whether the basic data has abnormal data or not;
if the abnormal data exist in the basic data, performing deviation rectification on each abnormal data;
and dividing the basic data after the deviation rectification processing into a plurality of data sets through a time sliding window with a preset length.
5. The method for building a prediction model of a demand for accessories according to claim 4, wherein the step of determining whether the abnormal data includes at least one of the first data and the second data and the basic data includes:
calculating a mean value and a standard deviation corresponding to the basic data, wherein the basic data comprises a plurality of reference data;
judging whether each reference data exceeds the sum of the mean value and the three standard deviations or whether each reference data is smaller than the difference of the mean value and the three standard deviations;
if any reference data exceeds the sum of the mean value and the three standard deviations, judging the corresponding reference data as the first data;
and if any reference data is smaller than the difference value between the mean value and the three standard deviations, judging the corresponding reference data as the second data.
6. The method for building a prediction model of a demand for accessories according to claim 5, wherein the step of performing error correction processing on each abnormal data includes:
judging whether the abnormal data is the first data or not;
if the abnormal data is the first data, passing a first formula
Assigning the first data as deviation correcting data;
if the abnormal data is the second data, passing a second formula
Assigning the first data to be the deviation correcting data;
wherein the content of the first and second substances,
for the assigned deviation correcting data, the deviation correcting data is calculated,
in order to be able to process the first data,
in order to be able to process the second data,
and the standard deviation corresponding to the basic data.
7. The method for building a prediction model of an accessory demand according to claim 1, wherein after the step of selecting a target prediction model with the highest prediction accuracy corresponding to each type of the accessory from all the reference prediction models, the method for building a prediction model of an accessory demand further comprises:
receiving a target type corresponding to an accessory input by a user and a target period to be predicted;
selecting the target prediction model corresponding to the target type as a standard prediction model;
acquiring the historical demand data corresponding to (i +1) continuous preset periods before the target period and the predicted value of the covariate data corresponding to the target period as reference data;
and inputting the reference data into the standard prediction model for prediction to obtain target demand data corresponding to the accessory of the target type.
8. An apparatus for building a prediction model of an accessory demand, the apparatus comprising:
the device comprises an acquisition module, a processing module and a processing module, wherein the acquisition module is used for acquiring basic data corresponding to accessories containing different reference types, and the basic data comprises historical demand data and covariate data which take a preset period as a unit;
the dividing module is used for dividing the basic data into a plurality of data sets through a time sliding window with preset length, wherein each data set comprises different characteristic factors, and each characteristic factor comprises the historical demand data corresponding to the (n-i) th preset period to the nth preset period and the covariate data corresponding to the (n +1) th preset period, wherein n and i are positive integers, and n > i;
the prediction module is used for performing regression prediction by using all the data sets as training sets through various machine learning algorithms to obtain reference prediction models corresponding to the machine learning algorithms;
and the selection module is used for selecting a target prediction model with the highest prediction precision corresponding to each type of accessory from all the reference prediction models.
9. A computer device, characterized in that it comprises a processor and a memory, said memory storing a computer program which, when executed on said processor, implements the method of predictive model construction of part requirements according to any one of claims 1 to 7.
10. A computer-readable storage medium, characterized in that it stores a computer program which, when executed on a processor, implements the method of predictive model construction of part requirements of any of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210104179.XA CN114118637A (en) | 2022-01-28 | 2022-01-28 | Method and device for building prediction model of accessory demand and computer equipment |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210104179.XA CN114118637A (en) | 2022-01-28 | 2022-01-28 | Method and device for building prediction model of accessory demand and computer equipment |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN114118637A true CN114118637A (en) | 2022-03-01 |
Family
ID=80361805
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210104179.XA Pending CN114118637A (en) | 2022-01-28 | 2022-01-28 | Method and device for building prediction model of accessory demand and computer equipment |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114118637A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN118011837A (en) * | 2024-04-08 | 2024-05-10 | 钛玛科（北京）工业科技有限公司 | Deviation rectifying control system based on MODBUS-RTU protocol communication |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112069809A (en) * | 2020-08-11 | 2020-12-11 | 桂林电子科技大学 | Missing text generation method and system |
| CN113887143A (en) * | 2021-10-21 | 2022-01-04 | 重庆邮电大学 | Spatial interpolation method and device for multi-source heterogeneous air pollutants and computer equipment |
| CN113962416A (en) * | 2020-07-02 | 2022-01-21 | 中科云谷科技有限公司 | Engineering machinery part stock prediction method, management method and system |
-
2022
- 2022-01-28 CN CN202210104179.XA patent/CN114118637A/en active Pending
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113962416A (en) * | 2020-07-02 | 2022-01-21 | 中科云谷科技有限公司 | Engineering machinery part stock prediction method, management method and system |
| CN112069809A (en) * | 2020-08-11 | 2020-12-11 | 桂林电子科技大学 | Missing text generation method and system |
| CN113887143A (en) * | 2021-10-21 | 2022-01-04 | 重庆邮电大学 | Spatial interpolation method and device for multi-source heterogeneous air pollutants and computer equipment |
Non-Patent Citations (1)
| Title |
|---|
| XIAOZHUANG SONG等: ""TINet: Multi-dimensional Traffic Data Imputation via Transformer Network"", 《ARTIFICIAL NEURAL NETWORKS AND MACHINE LEARNING – ICANN 2021 》 * |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN118011837A (en) * | 2024-04-08 | 2024-05-10 | 钛玛科（北京）工业科技有限公司 | Deviation rectifying control system based on MODBUS-RTU protocol communication |
| CN118011837B (en) * | 2024-04-08 | 2024-06-11 | 钛玛科（北京）工业科技有限公司 | Deviation rectifying control system based on MODBUS-RTU protocol communication |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US9336493B2 (en) | Systems and methods for clustering time series data based on forecast distributions | |
| Reiss et al. | Statistical analysis of extreme values | |
| Englberger et al. | Two-stage stochastic master production scheduling under demand uncertainty in a rolling planning environment | |
| CN111881158A (en) | Management report data processing method and device, computer system and readable storage medium | |
| Li et al. | Linear programming-based scenario reduction using transportation distance | |
| CN112036631B (en) | Purchasing quantity determining method, purchasing quantity determining device, purchasing quantity determining equipment and storage medium | |
| CN113361954A (en) | Attribution analysis method, attribution analysis device, attribution analysis equipment and storage medium | |
| Bladt et al. | Matrix representations of life insurance payments | |
| Wu et al. | A supervised learning-driven heuristic for solving the facility location and production planning problem | |
| CN114118637A (en) | Method and device for building prediction model of accessory demand and computer equipment | |
| Henckaerts et al. | When stakes are high: Balancing accuracy and transparency with Model-Agnostic Interpretable Data-driven suRRogates | |
| Butt et al. | Hybrid of deep learning and exponential smoothing for enhancing crime forecasting accuracy | |
| Martin et al. | Dynamical multi-parameter sizing of DDMRP buffers in finite capacity flow-shops | |
| CN110648045A (en) | Risk assessment method, electronic device and computer-readable storage medium | |
| Armaneri et al. | An integrated decision support approach for project investors in risky and uncertain environments | |
| Falini et al. | Spline based Hermite quasi-interpolation for univariate time series | |
| CN115481803A (en) | Financial time sequence prediction method, device and equipment based on industry crowding degree | |
| Rafiuzaman | Forecasting chaotic stock market data using time series data mining | |
| Wang et al. | Using incremental clustering technique in collaborative filtering data update | |
| CN114925275A (en) | Product recommendation method and device, computer equipment and storage medium | |
| Chern et al. | A heuristic master planning algorithm that includes fairness and flexibility | |
| CN111008749B (en) | Demand prediction method and device | |
| Korn | A simple method for modeling changes over time | |
| CN113837782A (en) | Method and device for optimizing periodic item parameters of time series model and computer equipment | |
| JP5789718B2 (en) | Risk scenario analysis system and risk scenario analysis 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: 20220301 |