GB2615205A - Long term precipitation prediction model establishing method, and long-term precipitation prediction method and apparatus - Google Patents
Long term precipitation prediction model establishing method, and long-term precipitation prediction method and apparatus Download PDFInfo
- Publication number
- GB2615205A GB2615205A GB2302648.7A GB202302648A GB2615205A GB 2615205 A GB2615205 A GB 2615205A GB 202302648 A GB202302648 A GB 202302648A GB 2615205 A GB2615205 A GB 2615205A
- Authority
- GB
- United Kingdom
- Prior art keywords
- month
- precipitation
- long
- precipitation amount
- prediction
- 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
- 238000001556 precipitation Methods 0.000 title claims abstract description 355
- 230000007774 longterm Effects 0.000 title claims abstract description 132
- 238000000034 method Methods 0.000 title claims abstract description 129
- 238000007637 random forest analysis Methods 0.000 claims abstract description 68
- 238000012216 screening Methods 0.000 claims abstract description 57
- 230000001419 dependent effect Effects 0.000 claims abstract description 24
- 238000012360 testing method Methods 0.000 claims abstract description 23
- 238000012549 training Methods 0.000 claims abstract description 7
- 238000010276 construction Methods 0.000 claims description 47
- 230000002860 competitive effect Effects 0.000 claims description 12
- 230000015654 memory Effects 0.000 claims description 12
- 238000004422 calculation algorithm Methods 0.000 claims description 6
- 238000010219 correlation analysis Methods 0.000 claims description 5
- 238000004891 communication Methods 0.000 claims description 2
- 238000003908 quality control method Methods 0.000 description 7
- 230000000694 effects Effects 0.000 description 6
- 238000010586 diagram Methods 0.000 description 5
- 238000012797 qualification Methods 0.000 description 5
- 238000004364 calculation method Methods 0.000 description 4
- 238000013473 artificial intelligence Methods 0.000 description 3
- 238000012545 processing Methods 0.000 description 3
- 238000004590 computer program Methods 0.000 description 2
- 238000012937 correction Methods 0.000 description 2
- 238000002790 cross-validation Methods 0.000 description 2
- 238000013500 data storage Methods 0.000 description 2
- 230000006870 function Effects 0.000 description 2
- 238000011160 research Methods 0.000 description 2
- 238000013179 statistical model Methods 0.000 description 2
- 238000013459 approach Methods 0.000 description 1
- 238000013528 artificial neural network Methods 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 238000004140 cleaning Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 230000003993 interaction Effects 0.000 description 1
- 238000010295 mobile communication Methods 0.000 description 1
- 238000012821 model calculation Methods 0.000 description 1
- 230000010355 oscillation Effects 0.000 description 1
- 230000011664 signaling Effects 0.000 description 1
- 238000012706 support-vector machine Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01W—METEOROLOGY
- G01W1/00—Meteorology
- G01W1/10—Devices for predicting weather conditions
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01W—METEOROLOGY
- G01W1/00—Meteorology
- G01W1/18—Testing or calibrating meteorological apparatus
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N20/00—Machine learning
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N20/00—Machine learning
- G06N20/20—Ensemble learning
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N5/00—Computing arrangements using knowledge-based models
- G06N5/01—Dynamic search techniques; Heuristics; Dynamic trees; Branch-and-bound
-
- 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
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/10—Services
- G06Q50/26—Government or public services
-
- 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
- Y02A—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE
- Y02A90/00—Technologies having an indirect contribution to adaptation to climate change
- Y02A90/10—Information and communication technologies [ICT] supporting adaptation to climate change, e.g. for weather forecasting or climate simulation
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Business, Economics & Management (AREA)
- Software Systems (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Evolutionary Computation (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Mathematical Physics (AREA)
- Artificial Intelligence (AREA)
- Data Mining & Analysis (AREA)
- Environmental & Geological Engineering (AREA)
- Tourism & Hospitality (AREA)
- Economics (AREA)
- Strategic Management (AREA)
- Human Resources & Organizations (AREA)
- Development Economics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Medical Informatics (AREA)
- Marketing (AREA)
- General Business, Economics & Management (AREA)
- Environmental Sciences (AREA)
- Biodiversity & Conservation Biology (AREA)
- Atmospheric Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Ecology (AREA)
- Computational Linguistics (AREA)
- General Health & Medical Sciences (AREA)
- Quality & Reliability (AREA)
- Operations Research (AREA)
- Entrepreneurship & Innovation (AREA)
- Game Theory and Decision Science (AREA)
- Primary Health Care (AREA)
- Health & Medical Sciences (AREA)
- Educational Administration (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
A long-term precipitation prediction model establishing method, and a long-term precipitation prediction method and apparatus. The establishing method comprises: acquiring a sample set (step S101); on the basis of a false discovery rate control method for multiple hypothesis testing and a random forest model, screening interpretation variables in the sample set to obtain a prediction factor affecting the precipitation of a corresponding month of the following year (step S102); according to the prediction factor affecting the precipitation amount of the corresponding month of the following year, and the precipitation amount of the corresponding month of the following year, performing random forest modeling, and training to obtain a long-term precipitation prediction model for the precipitation amount of the corresponding month (step S103). By means of the present long-term precipitation prediction model establishing method, and the long-term precipitation prediction method and apparatus, variable screening can be optimized from an experience-dependent method to a data-dependent method, the problem of false positive error rate of the random forest method during empirical screening variables can be improved, and the accuracy and reliability of model prediction can be effectively improved.
Description
LONG TERM PRECIPITATION PREDICTION MODEL ESTABLISHING METHOD, AND LONG-TERM PRECIPITATION PREDICTION METHOD AND APPARATUS
TECHNICAL FIELD
100011 The present invention relates to the technical field of long-term hydrology, in particular to a construction method for a long-term precipitation prediction model, and a longterm precipitation prediction method and apparatus.
BACKGROUND
[0002] The long-term precipitation quantitative prediction methods can be divided into a dynamic numerical method, a mathematical statistics method, and an artificial intelligence method. In the dynamic numerical method, precipitation prediction is performed by simulating future weather conditions by means of a maritime thermal dynamic model. The physical CO mechanism thereof is clear, but the model calculation is complex, and there is still a large CN deviation in the precipitation prediction in a local area. In the mathematical statistics method, CO from the perspective of statistics, the future precipitation is inferred by constructing the statistical relationship between climate factors and precipitation, or analyzing the characteristics (such as trend and period) of a precipitation sequence itself As the correlative relation between precipitation and previous climate factors hardly satisfies the assumed conditions of the statistical model, it is difficult for the statistical model to accurately predict real-time changes of precipitation. With the development of information technology, artificial intelligence methods such as artificial neural network, support vector machine and random forest provide new ideas for precipitation prediction. It can effectively excavate the correlation between the characteristic factors and the prediction quantity by selecting appropriate predictive factors from the influence factors such as climate and sea temperature and constructing models to simulate the relationship between precipitation and prediction characteristic factors. At present, it has been widely used in long-term precipitation prediction.
[00031 The key steps of the long-term precipitation prediction method based on artificial intelligence are the screening of prediction factors. Currently, the factor screening methods mainly include a correlation coefficient method, a stepwise regression method, and a full subset regression method. However, the described method only considers a correlation between a single factor and a prediction target; and a random forest algorithm can consider the importance of the single factor, and can consider a complex association between factors, and therefore is widely applied to mid-long-term prediction. However, there is a certain false-positive false rate problem in the variables selected by the random forest algorithm, which will lead to the accuracy and reliability of the model difficult to meet the needs.
SUMMARY
[0004] In view of this, embodiments of the present invention provide a construction method for a long-term precipitation prediction model, and a long-term precipitation prediction method and apparatus, so as to solve the technical problem of low accuracy when using random forest for precipitation prediction in the prior art.
[0005] The technical solution provided by the present invention is as follows: [0006] A first aspect of embodiments of the present invention provides a construction method for a long-term precipitation prediction model. The construction method comprises: acquiring a sample set, wherein the sample set comprises explanatory variables and dependent variables, each of the explanatory variables comprises a precipitation amount of each month of a current year and an observation value of each month of the current year for a key climate factor affecting the precipitation amount of the corresponding month, and each of the dependent variables comprises a precipitation amount of the corresponding month of a next year; screening the explanatory variables on the basis of a false discovery rate control method of a multi-hypothesis test and a random forest model, so as to obtain a prediction factor affecting the precipitation amount of the corresponding month of the next year; and performing random forest modeling according to the prediction factor affecting the precipitation amount of the corresponding month of the next year and the precipitation amount of the corresponding month of the next year, and training the model, so as to obtain a long-term precipitation prediction model for the precipitation amount of the corresponding month.
[0007] Optionally, the step of acquiring a sample set comprises: acquiring historical precipitation data and climate factor data; determining, according to the historical precipitation data and the climate factor data, the precipitation amount of each month and the key climate factor affecting the precipitation amount of the corresponding month; determining, according to a key factor affecting the precipitation amount of the corresponding month, an observation value of each month of the current year for the key climate factor affecting the precipitation amount of the corresponding month; and constructing a sample set according to the precipitation amount of each month of the current year, the observation value of each month of the current year for the key climate factor affecting the precipitation amount of the corresponding month, and the precipitation amount of the corresponding month of the next year.
100081 Optionally, after acquiring historical precipitation data and climate factor data, the method further comprises: correcting and quality controlling the historical precipitation data and the climate factor data; determining, according to the historical precipitation data and the climate factor data, the precipitation amount of each month and the key climate factor affecting the precipitation amount of the corresponding month, comprising: screening the climate factor data on the basis of a grey correlation analysis algorithm, so as to obtain the key climate factor affecting the precipitation amount of the corresponding month.
[0009] Optionally, the step of screening the explanatory variables on the basis of a false discovery rate control method of a multi-hypothesis test and a random forest model so as to obtain a prediction factor affecting the precipitation amount of the corresponding month of the next year comprises. constructing a knockoff variable of each of the explanatory variables; scoring the explanatory variables and the knockoff variables according to a random forest model, so as to obtain explanatory variable scores and knockoff variable scores; and screening the interpretation variable scores and the knockoff variable scores, on the basis of the false discovery rate control method so as to obtain a prediction factor affecting the precipitation amount of the corresponding month of the next year [0010] Optionally, the step of screening the interpretation variable scores and the knockoff variable scores on the basis of the false discovery rate control method so as to obtain a prediction factor affecting the precipitation amount of the corresponding month of the next year comprises: calculating the explanatory variable scores and the knockoff variable scores, so as to obtain a competitive scoring of each of the explanatory variables; and screening the competitive scoring on the basis of a pre-set false rate level, so as to obtain a prediction factor affecting the precipitation amount of the corresponding month of the year.
100111 A second aspect of embodiments of the present invention provides a long-term precipitation prediction method. The method comprises: acquiring a precipitation amount of each month of a previous year of the month to be predicted and an observation value of each month of the previous year for a key climate factor affecting the precipitation amount of the month to be predicted; and inputting the precipitation amount of each month of a previous year and the observation value of each month of the previous year for the key climate factor affecting the precipitation amount of the month to be predicted into the long-term precipitation prediction model of the month to be predicted constructed by the construction method for a long-term precipitation prediction model described in the first aspect and any one of the first aspect of the embodiments of the present invention, so as to obtain the precipitation amount of the month to be predicted.
[0012] A third aspect of embodiments of the present invention provides a construction apparatus for a long-term precipitation prediction model. The apparatus comprises: a sample acquisition module, wherein the sample acquisition module is used for acquiring a sample set, the sample set comprises explanatory variables and dependent variables, each of the explanatory variables comprises a precipitation amount of each month of a current year and an observation value of each month of the current year for a key climate factor affecting the precipitation amount of the corresponding month, and each of the dependent variables comprises a precipitation amount of the corresponding month of a next year; a screening module, wherein the screening module is used for screening the explanatory variables on the basis of a false discovery rate control method of a multi-hypothesis test and a random forest model, so as to obtain a prediction factor affecting the precipitation amount of the corresponding month of the next year; and a modeling module, wherein the modeling module is used for performing random forest modeling according to the prediction factor affecting the precipitation amount of the corresponding month of the next year and the precipitation amount of the corresponding month of the next year, and training the model, so as to obtain a long-term precipitation prediction model for the precipitation amount of the corresponding month.
[0013] A fourth aspect of embodiments of the present invention provides a long-term precipitation prediction apparatus. The apparatus comprises: a data acquisition module, wherein the data acquisition module is used for acquiring a precipitation amount of each month of a previous year of the month to be predicted and an observation value of each month of the previous year for a key climate factor affecting the precipitation amount of the month to be predicted; and a prediction module, wherein the prediction module is used for inputting the precipitation amount of each month of a previous year and the observation value of each month of the previous year for the key climate factor affecting the precipitation amount of the month to be predicted into the long-term precipitation prediction model of the month to be predicted constructed by the construction method for a long-term precipitation prediction model described in the first aspect and any one of the first aspect of the embodiments of the present invention, so as to obtain the precipitation amount of the month to be predicted.
100141 A fifth aspect of embodiments of the present invention provides a computer readable storage medium having a computer instruction stored thereon, wherein the computer instruction is used to enable a computer to execute the construction method for a long-term precipitation prediction model described in the first aspect and any one of the first aspect of the embodiments of the present invention and the long-term precipitation prediction method described in the second aspect of the embodiments of the present invention.
[0015] A sixth aspect of embodiments of the present invention provides an electronic device, comprising a memory and a processor, wherein the memory having a computer instruction stored thereon and the processor are in communication connection with each other, and when executing the computer instruction, the processor executes the construction method for a long-term precipitation prediction model described in the first aspect and any one of the first aspect of the embodiments of the present invention and the long-term precipitation prediction method described in the second aspect of the embodiments of the present invention.
[0016] The technical solution provided by the present invention has the following effects: [0017] In the construction method and apparatus for a long-term precipitation prediction model provided in the embodiments of the present invention, first obtaining, as a sample set, data such as the precipitation amount of each month of the current year, the observation value of each month of the current year for the key climate factor affecting the precipitation amount of the corresponding month, and the precipitation amount of the corresponding month of the next year, then screening a prediction factor by using a random forest in combination with a false discovery rate control method of multi-hypothesis tests; finally, performing random forest modeling by using a prediction factor obtained by screening, so as to obtain a long-term precipitation prediction model of a corresponding month. Hence, the construction method for a long-term precipitation prediction model in combination with a false discovery rate control method of a statistical multi-hypothesis test effectively performs quality control on the variable screening in the random forest model; and the method can optimize variable screening from an experience-dependent method to a data-dependent method, can improve a false-positive false rate problem of a random forest method when empirically screening variables, and can effectively improve the accuracy and reliability of model prediction.
[0018] In the long-term precipitation prediction method and apparatus provided in the embodiments of the present invention, the long-term precipitation prediction model of a corresponding month constructed by using the construction method for the described long-term precipitation prediction model is used to perform precipitation prediction of a month to be predicted As the constructed long-term precipitation prediction model is constructed by using a false discovery rate control method of a multi-hypothesis test in combination with a random forest model, the prediction accuracy and reliability are relatively high. Hence, the long-term precipitation prediction method has a relatively high prediction accuracy.
BRIEF DESCRIPTION OF THE DRAWINGS
[0019] To describe the technical solutions in the embodiments of the present invention or in the prior art more clearly, the following briefly introduces the drawings required for describing the embodiments or the prior art. Apparently, the drawings in the following description show some embodiments of the present invention, other drawings can also be obtained according to these drawings without creative efforts.
[0020] Fig. 1 is a flowchart of a construction method for a long-term precipitation prediction model according to an embodiment of the present invention; [0021] Fig 2 is a flowchart of a construction method for a long-term precipitation prediction model according to another embodiment of the present invention, [0022] Fig 3 is a flowchart of a construction method for a long-term precipitation prediction model according to another embodiment of the present invention, [0023] Fig 4 is a flowchart of a construction method for a long-term precipitation prediction model according to another embodiment of the present invention, [0024] Fig 5 is a flowchart of a long-term precipitation prediction method according to an embodiment of the present invention; [0025] Fig. 6(a) to Fig. 6(c) are schematic diagrams of prediction results of a long-term precipitation prediction method according to an embodiment of the present invention; [0026] Fig. 7 is a structural block diagram of a construction apparatus for a long-term precipitation prediction model according to an embodiment of the present invention; [0027] Fig. 8 is a structural block diagram of a long-term precipitation prediction apparatus according to an embodiment of the present invention; [0028] Fig. 9 is a schematic structural diagram of a computer readable storage medium according to an embodiment of the present invention, and [0029] Fig. 10 is a schematic structural diagram of an electronic device according to an embodiment of the present invention.
DETAILED DESCRIPTION OF THE EMBODIMENTS
[0030] To make a person skilled in the art better understand the solutions of the present invention, the following clearly and completely describes the technical solutions in the embodiments of the present invention with reference to the drawings in the embodiments of the present invention. Apparently, the described embodiments are merely a part rather than all of the embodiments of the present invention. All other embodiments obtained by a person of ordinary skill in the art on the basis of the embodiments of the present invention without creative efforts shall belong to the scope of protection of the present invention.
[0031] The terms "first", "second", "third", "fourth" and the like in the description, claims, and the drawings of the present invention are used to distinguish similar objects, and are not necessarily used to describe a specific sequence or order. It should be understood that the data so used can be interchanged where appropriate so that the embodiments described herein can be practiced in an order other than that illustrated or described herein. In addition, the terms "comprise" and "have", and any variations thereof, are intended to cover a non-exclusive inclusion, for example, a process, method, system, product, or device that comprises a series of steps or units is not necessarily limited to those steps or units that are expressly listed, but can comprise other steps or units that are not expressly listed or inherent to such process, method, product, or device.
[0032] As described in the background, a random forest algorithm does not have a variable screening function, and a traditional approach is to empirically select some of variables according to the importance score of variables given by the random forest or the out of pocket error value, and then remodel the selected variables. Therefore, variables selected depending on experience may not exist or have a weak correlation with a target to be predicted, that is, the selected variables have a certain false-positive false rate problem, which will lead to the accuracy and reliability of the model difficult to meet the needs.
[0033] In view of this, embodiments of the present invention provide a construction method for a long-term precipitation prediction model and a long-term precipitation prediction method. By effectively performing quality control on variable screening in a random forest model in combination with a false discovery rate control method of multi-hypothesis test, the method can improve a false-positive false rate problem of a random forest method when empirically screening variables, and can effectively improve the accuracy and reliability of model prediction.
[0034] According to the embodiments of the present invention, a construction method for a long-term precipitation prediction model and a long-term precipitation prediction method and apparatus are provided. It should be noted that the steps shown in the flowchart of the drawings can be executed in a computer system such as a group of computer executable instructions. Furthermore, although a logic sequence is shown in the flowchart, in some cases, the shown or described steps can be executed in a sequence different from that described herein.
[0035] The present embodiment provides a construction method for a long-term precipitation prediction model, which can be used for an electronic device, such as a computer, a mobile phone, and a tablet computer. Fig. 1 is a flowchart of a construction method for a long-term precipitation prediction model according to an embodiment of the present invention. As shown in Fig. 1, the method comprises the following steps: [0036] Step 5101 acquiring a sample set, wherein the sample set comprises explanatory variables and dependent variables, each of the explanatory variables comprises a precipitation amount of each month of a current year and an observation value of each month of the current year for a key climate factor affecting the precipitation amount of the corresponding month, and each of the dependent variables comprises a precipitation amount of the corresponding month of a next year.
[0037] Specifically, when constructing a long-term precipitation prediction model, precipitation data and climate factor data of a current year, and precipitation data of a next year are used as a sample set. There may be a certain difference in the precipitation amount of each month, and therefore a long-term precipitation prediction model can be constructed for each month. For example, if a long-term precipitation prediction model of January is constructed, using the precipitation data of 12 months of the current year, the key climate factor data and the precipitation data of January of the next year as a sample set of January; if a long-term precipitation prediction model of February is constructed, using the precipitation data and climate factor data of 12 months of the current year arid the precipitation data of February of the next year as a sample set of February; by analogy, a long-term precipitation prediction model of 12 months is constructed by using 12 sample sets.
[0038] In one embodiment, when determining a key climate factor, screening can be performed from known climate factors to determine a key climate factor affecting the precipitation amount of each month. The known climate factors can comprise 130 climate system indexes provided by the National Climate Center of the Chinese Meteorological Administration, comprising 88 atmospheric circulation indexes such as the subtropical high area index in the Northern Hemisphere and the subtropical high area index in North Africa; 26 sea temperature indexes, such as a sea surface temperature anomaly index of NINO zones 1 + 2 and a sea surface temperature anomaly index of NINO zone 3; and 16 indexes such as a solar black index and a southern oscillation index. In addition, the known climate factors can also comprise other climate factor data, which is not limited in the embodiments of the present invention.
[0039] Specifically, when determining key climate factor data affecting the precipitation amount of each month, 20 key climate factors can be screened out from 130 climate system indexes by using grey correlation analysis as a key climate factor affecting the precipitation amount of each month. The key climate factor affecting the precipitation amount of the corresponding month are determined on the basis of actual conditions of each month, and thus the key climate factor affecting the precipitation amount of each month may be the same or may be different. In one embodiment, when a key climate factor affecting a corresponding month is determined, it can be determined according to the month of a long-term precipitation prediction model that can be constructed, for example, when a long-term precipitation prediction model of January is constructed, 20 key climate factors affecting January are determined first, and then the observation values of the 20 key climate factors in the 12 months of the current year are determined. Hence, the explanatory variables in the sample set of long-term precipitation prediction model of January comprise the precipitation amount of 12 months of the current year and the observation values of 12 months of the current year affecting the key climate factors of January.
100401 Step S102: screening the explanatory variables on the basis of a false discovery rate control method of a multi-hypothesis test and a random forest model, so as to obtain a prediction factor affecting the precipitation amount of the corresponding month of the next year. Specifically, on the basis of the described contents, it can be determined that when a sample set of a long-term precipitation prediction model is constructed, the precipitation data and climate factor data of 12 months of a current year are used as explanatory variables, and there are a large number of explanatory variables in one sample set; therefore, the explanatory variables can be screened so as to obtain a prediction factor affecting the precipitation amount of the corresponding month of the next year. The prediction factor is an explanatory variable that has a great influence on the precipitation amount of the corresponding month of the next year in the precipitation data and climate factor data of 12 months of the current year. As the long-term precipitation prediction model needs to be constructed every month, the explanatory variables in the sample set required by each long-term precipitation prediction model can be screened separately. [0041] In one embodiment, when performing explanatory variable screening, a random forest in combination with a false discovery rate control method of a multi-hypothesis test is used to screen so as to obtain a prediction factor having a greater influence on the precipitation amount of the corresponding month of the next year. The predictor therein comprises precipitation data for 12 months of the current year and partial data of the key climate factor in observation values of each month of the current year.
100421 Step S103: performing random forest modeling according to the prediction factor affecting the precipitation amount of the corresponding month of the next year and the precipitation amount of the corresponding month of the next year, so as to obtain a long-term precipitation prediction model for the precipitation amount of the corresponding month. Specifically, after screening the prediction factor of the precipitation amount of the corresponding month of the next year, the random forest modeling can be performed for the prediction factor and the precipitation amount of the corresponding month of the next year, so as to obtain the long-term precipitation prediction model of the precipitation amount of the corresponding month. For example, when constructing a long-term precipitation prediction model of January, random forest modeling is performed by using a screened prediction factor affecting the precipitation amount of the next month and the precipitation amount of the next month, so as to obtain a long-term precipitation prediction model of January. Similarly, long-term precipitation predictive models for other months are constructed in the same manner. When performing the random forest modeling, an existing random forest modeling method is used, and it will not be described again herein.
100431 In the construction method for a long-term precipitation prediction model provided in the embodiments of the present invention, first obtaining, as a sample set, data such as the precipitation amount of each month of the current year, the observation value of each month of the current year for the key climate factor affecting the precipitation amount of the corresponding month, and the precipitation amount of the corresponding month of the next year; then screening a prediction factor by using a random forest in combination with a false discovery rate control method of multi-hypothesis tests; finally, performing random forest modeling by using a prediction factor obtained by screening, so as to obtain a long-term precipitation prediction model of a corresponding month. Hence, the construction method for a long-term precipitation prediction model in combination with a false discovery rate control method of a statistical multi-hypothesis test effectively performs quality control on the variable screening in the random forest model; and the method can optimize variable screening from an experience-dependent method to a data-dependent method, can improve a false-positive false rate problem of a random forest method when empirically screening variables, and can effectively improve the accuracy and reliability of model prediction.
[0044] In one embodiment, as shown in Fig 2, acquiring the sample set specifically comprises the following steps: [0045] Step S201: acquiring historical precipitation data and climate factor data. Specifically, in order to make the modeling more accurate, when corresponding data is acquired, historical precipitation data and other climate factor data of a corresponding research region can be acquired. For example, historical precipitation data and climate factor data can be acquired for a region.
[0046] In one embodiment, after obtaining the data, the historical precipitation data and climate factor data are corrected and quality-controlled, so that the constructed model is more accurate. The data correction and quality control specifically comprise merging daily precipitation data into monthly precipitation data, performing format conversion on the monthly precipitation data, cleaning missing data and performing data standardization, and the like. Processing of missing data: for example, for the data with less than 3 days of missing precipitation of each month, the data of adjacent dates are directly used for linear interpolation, and then the monthly data are obtained; if the data of the current month with more than 3 days of precipitation is missing, the historical average precipitation of the current month shall be used for interpolation.
[0047] In particular, the climate factor data comprises 130 climate system indexes provided by the National Climate Center of the Chinese Meteorological Administration mentioned above, and can also comprise other climate factor data affecting the precipitation amount.
[0048] Step S202: determining, according to the historical precipitation data and the climate factor data, the precipitation amount of each month and the key climate factor affecting the precipitation amount of the corresponding month. Specifically, when a key climate factor affecting the precipitation amount of the corresponding month is determined, the climate factor data is screened by using a gray correlation analysis algorithm, so as to obtain the key climate factor affecting the precipitation amount of the corresponding month. During screening, the correlation between each climate factor and the precipitation amount of each month is calculated, so as to screen and obtain the first 20 key climate factors with high correlation degree with precipitation of different months. It should be noted that, other numbers such as the first 10 or the first 30 key climate factors can also be screened. The embodiment of the present invention does not limit the number of the key climate factors obtained by screening. [0049] Step S203: determining, according to a key factor affecting the precipitation amount of the corresponding month, an observation value of each month of the current year for the key climate factor affecting the precipitation amount of the corresponding month. Specifically, after a key climate factor affecting the precipitation amount of a corresponding month is determined, an observation value of each month of the current year for the key climate factor is further determined as an explanatory variable. For example, when a long-term precipitation prediction model of January is constructed, 20 key climate factors affecting January are determined first, and then the observation values of the 20 key climate factors in the 12 months of the current year are determined.
[0050] Step S204: constructing a sample set according to the precipitation amount of each month of the current year, the observation value of each month of the current year for the key climate factor affecting the precipitation amount of the corresponding month, and the precipitation amount of the corresponding month of the next year. Specifically, when a sample set is constructed, different sample sets are constructed on the basis of different months. For example, if a sample set of January is constructed, the precipitation amount of 12 months of the current year, the observation value of 12 months of the current year for the key climate factor affecting the precipitation amount of January, and the precipitation amount of January of the next year are used to construct a sample set. Hence, 12 sample sets can be constructed.
[0051] In one embodiment, the step of screening the explanatory variables on the basis of a false discovery rate control method of a multi-hypothesis test and a random forest model so as to obtain a prediction factor affecting the precipitation amount of the corresponding month of the next year comprises the following steps: [0052] Step S301: constructing a knockoff variable for each of the explanatory variables. Specifically, for each sample set, if 20 key climate factors affecting a precipitation amount of each month are selected, there are 240 observation values of the 20 key climate factors of each month from January to December. In addition, a precipitation amount of each month from January to December is also comprised, and thus, each sample set comprises 252 explanatory variables in total. Then, on the basis of the knockoff method, knockoff variables) of 252 explanatory variables in each sample set can be calculated. For a specific calculation manner of the knockoff variable, an existing knockoff calculation method can be used, and details are not repeatedly described herein [0053] Step S302: scoring the explanatory variables and the knockoff variables according to a random forest model, so as to obtain explanatory variable scores and knockoff variable scores; specifically, the calculated knockoff variables and the corresponding explanatory variables of each of the explanatory variables can be input into the random forest model for the scoring of factor importance, so as to obtain a score 2 of each of the explanatory variables and a score of each of the knockoff variables. The importance scoring is performed by using a random forest model, and calculation can be performed by using a Gini index or an Out of Bag (00B) false rate, which is not limited in the embodiment of the present invention.
[0054] Step S303: screening the interpretation variable scores and the knockoff variable scores on the basis of the false discovery rate control method so as to obtain a prediction factor affecting the precipitation amount of the corresponding month of the next year. Specifically, for the score 2 of each of the explanatory variables and the score of each of the corresponding knockoff variables, a competitive score of each of the explanatory variable can be calculated by means of a difference value, and the competitive score is represented by the following formula: After a competitive score of each of the interpretation variables is calculated, a false discovery rate (FDR) control method is used to screen the interpretation variables according to the competitive score.
100551 In one embodiment, when the explanatory variables are screened according to the competitive scoring, the following formula is mainly used to screen the explanatory variables of which the competitive scoring is higher than the scoring threshold T [0056] [0057] wherein, and a denotes a pre-set false rate level. If all explanatory variables whose final scores are higher than T are selected as the final prediction factors, the false discovery rate FDR of the screening factors can be strictly controlled to be < a.
[0058] In one embodiment, the construction method for a long-term precipitation prediction model is specifically implemented by using the following procedure: [0059] Step S401: downloading, from a public website, daily precipitation data of the corresponding research region from 1980 to 2020, wherein the unit of the precipitation data is mm.
[0060] Step S402: performing correction and quality control on the daily precipitation data, so as to obtain a precipitation amount of each month. For the data with less than 3 days of missing precipitation of each month, the data of adjacent dates are directly used for linear interpolation, and then the monthly data are obtained; if the data of the current month with more than 3 days of precipitation is missing, the historical average precipitation of the current month shall be used for interpolation. [0061] Step S403: determining a key climate factor for the precipitation amount of each month on the basis of the effect of the climate factor on the precipitation amount of each month. Specifically, with regard to the precipitation amount of 41 years from 1980 to 2020, first 20 climate factors having a high correlation with the current month are acquired month by month from known climate factors by using grey correlation analysis, and are taken as key climate factors of the precipitation amount of each month.
[0062] Step S404: constructing a sample set on the basis of the precipitation amount of each month and the key climate factor affecting the precipitation amount of each month. Specifically, taking a sample set required for modeling of January as an example, a sample set is constructed by using the obtained observation values of the precipitation amount of each month from January to December in 1980 and the key climate factor of each month in 1980 affecting the precipitation amount of January as explanatory variables, and using the precipitation amount of January in 1981 as a dependent variable. A sample set is constructed by using the observation values of the precipitation amount of each month from January to December in 1981 and the key climate factor of each month in 1981 affecting the precipitation amount of January as explanatory variables, and using the precipitation amount of January in 1982 as a dependent variable. By analogy, a total of 40 sample sets required for modeling of January can be obtained. In the same manner, 12 sample sets required for modeling of each month of other months can be obtained. When 20 key climate factors are obtained by screening, a total of 252 explanatory variables and 1 dependent variable are comprised in each sample set [0063] Step S405: screening explanatory variables in a sample set so as to obtain a prediction factor affecting the precipitation amount of the corresponding month of the next year. Specifically, for each sample set, knockoff variables thereof are first constructed respectively for its 252 explanatory variables by using a knockoff method, then using a random forest model to score the importance of all the original explanatory variables and corresponding knockoff variables, then obtaining competitive scores of 252 variables, setting a given FDR threshold value as 0.1, and calculating a required scoring threshold value, and finally screening the explanatory variables with a competitive score above a scoring threshold as prediction factors.
[0064] Step S406: re-performing random forest modeling by using the prediction factor obtained by screening and the precipitation amount of the next year. Specifically, in step S401, an explanatory variable can be screened to obtain a required prediction factor for each sample set, and the sample set can be reconstructed by using the prediction factor and a corresponding dependent variable. For example, with regard to the construction of a long-term precipitation prediction model of January, the described 40 sample sets are all reconstructed in the described manner, and then random forest modeling is performed by using the reconstructed 40 sample sets so as to obtain a long-term precipitation prediction model of January. In the same manner, long-term precipitation prediction models for other months can be obtained. [0065] An embodiment of the present invention further provides a long-term precipitation prediction method. As shown in Fig. 5, the method comprises the following steps: [0066] Step S501: acquiring a precipitation amount of each month of a previous year of the month to be predicted and an observation value of each month of the previous year for a key climate factor affecting the precipitation amount of the month to be predicted. Specifically, when a long-term precipitation prediction model constructed by using the described construction method for a long-term precipitation prediction model is used to perform precipitation prediction, it is necessary to first determine a month to be predicted, and then determine a long-term precipitation prediction model of the corresponding month. For example, if it is required to predict precipitation of February in 2022, it is necessary to obtain a long-term precipitation prediction model of February. Then, the precipitation amount of each month from January to December in 2021 and the observation values of 12 months in 2021 for the key climate factor affecting the precipitation amount of February are acquired, wherein the method for determining the key climate factor affecting the precipitation amount of January uses the method for determining the key climate factor in the construction method for a long-term precipitation prediction model, which will not be repeated herein.
[0067] Step S502: inputting the precipitation amount of each month of a previous year and the observation value of each month of the previous year for the key climate factor affecting the precipitation amount of the month to be predicted into the long-term precipitation prediction model of the month to be predicted constructed by the construction method for a long-term precipitation prediction model described in any one of the foregoing embodiments, so as to obtain the precipitation amount of the month to be predicted. Specifically, by taking the predicted precipitation amount of February in 2022 as an example, the acquired precipitation amount of each month from January to December in 2021 and the observation value of January to December in 2021 for the key climate factor affecting the precipitation amount of February are input into the long-term precipitation prediction model of February, so as to output and obtain the precipitation amount of February in 2022.
[0068] In the long-term precipitation prediction method provided in the embodiments of the present invention, the long-term precipitation prediction model of a corresponding month constructed by using the construction method for the described long-term precipitation prediction model is used to perform precipitation prediction of a month to be predicted. As the constructed long-term precipitation prediction model is constructed by using a false discovery rate control method of a multi-hypothesis test in combination with a random forest model, the prediction accuracy and reliability are relatively high. Hence, the long-term precipitation prediction method has a relatively high prediction accuracy.
[0069] In one embodiment, he accuracy of the long-term precipitation prediction method can be evaluated by comparison. Considering that a mid-long term quantitative precipitation prediction error is qualified within a 20% range of a multi-year variation, a qualified rate (P), a root mean square error (RMSE) and a mean absolute error (MAE) are respectively used to evaluate the accuracy of quantitative precipitation prediction. The specific calculation formula is as follows: [0071] [0072] wherein h(x(p) is the prediction data of the ith sample point, and y(i) is the real data of the ith sample point. As 40 samples are used herein, a method of 10-fold cross validation is used herein to verify the model effect, that is, 40 samples are divided into 10 groups, 9 groups of samples (36) are sequentially selected as the training set, and the remaining 1 group of samples (4) is used as the test set, which is repeated 10 times and all the samples are traversed. The final demonstrated model effects are the average of 10 cross validation results.
[0073] In order to perform the comparison of methods, a random forest method is also used, wherein the random forest method directly scores the 252 explanatory variables for importance in the described steps, and screens the first five explanatory variables for the importance scoring as prediction factors, and the remaining steps are consistent.
[0074] The final effects are as shown in Fig. 6(a), Fig. 6(b) and Fig. 6(c). The model prediction qualification rate of the random forest model on the basis of the knockoff method from January to December, except April, is not smaller than that of the simple random forest method. Among them, the qualification rate of August is the highest, reaching 97.5%, while the random forest method is 95%; in May, the qualification rate of random forest method on the basis of a knockoff method is 90%, while that of random forest method was only 77.5%; the precipitation in April and November is less, and the qualification rate is low, 52.5% and 57.5% respectively, while that in random forests is 57.5% and 55%. For a qualification rate from January to December, the random forest model on the basis of the knockoff method is 72.5% and the simple random forest method is only 66.7%. The results of RMSE and MAE show that the fitting errors of the random forest model from January to December on the basis of the knockoff method are substantially lower than the pure random forest method. Therefore, the random forest method in combination with hypothesis test for variable selection quality control is significantly better than the traditional random forest method.
[0075] An embodiment of the present invention further provides a construction apparatus for a long-term precipitation prediction model. As shown in Fig. 7, the apparatus comprises: [0076] a sample acquisition module, wherein the sample acquisition module is used for acquiring a sample set, the sample set comprises explanatory variables and dependent variables, each of the explanatory variables comprises a precipitation amount of each month of a current year and an observation value of each month of the current year for a key climate factor affecting the precipitation amount of the corresponding month, and each of the dependent variable comprises a precipitation amount of the corresponding month of a next year; for specific content, reference can be made to a corresponding part of the foregoing method embodiment, and details are not repeatedly described herein; [0077] a screening module, wherein the screening module is used for screening the explanatory variables on the basis of a false discovery rate control method of a multi-hypothesis test and a random forest model, so as to obtain a prediction factor affecting the precipitation amount of the corresponding month of the next year; for specific content, reference can be made to a corresponding part of the foregoing method embodiment, and details are not repeatedly described herein; and [0078] a modeling module, wherein the modeling module is used for performing random forest modeling according to the prediction factor affecting the precipitation amount of the corresponding month of the next year and the precipitation amount of the corresponding month of the next year, and training the model, so as to obtain a long-term precipitation prediction model for the precipitation amount of the corresponding month; for specific content, reference can be made to a corresponding part of the foregoing method embodiment, and details are not repeatedly described herein.
[0079] In the construction apparatus for a long-term precipitation prediction model provided in the embodiments of the present invention, first obtaining, as a sample set, data such as the precipitation amount of each month of the current year, the observation value of each month of the current year for the key climate factor affecting the precipitation amount of the corresponding month, and the precipitation amount of the corresponding month of the next year; then screening a prediction factor by using a random forest in combination with a false discovery rate control method of multi-hypothesis tests; finally, performing random forest modeling by using a prediction factor obtained by screening, so as to obtain a long-term precipitation prediction model of a corresponding month. Hence, the construction apparatus for a long-term precipitation prediction model in combination with a false discovery rate control method of a multi-hypothesis test effectively performs quality control on the variable screening in the random forest model; and the apparatus can optimize variable screening from an experience-dependent method to a data-dependent method, can improve a false-positive false rate problem of a random forest method when empirically screening variables, and can effectively improve the accuracy and reliability of model prediction.
[0080] For the functional description of the construction apparatus for a long-term precipitation prediction model provided by the embodiment of the present invention, reference can be made to the description of the construction method for a long-term precipitation prediction model in the foregoing embodiment.
[0081] An embodiment of the present invention further provides a long-term precipitation prediction apparatus. As shown in Fig. 8, the apparatus comprises: [0082] a data acquisition module, wherein the data acquisition module is used for acquiring a precipitation amount of each month of a previous year of the month to be predicted and an observation value of each month of the previous year for a key climate factor affecting the precipitation amount of the month to be predicted; for specific content, reference can be made to a corresponding part of the foregoing method embodiment, and details are not repeatedly described herein; and [0083] a prediction module, wherein the prediction module is used for inputting the precipitation amount of each month of a previous year and the observation value of each month of the previous year for the key climate factor affecting the precipitation amount of the month to be predicted into the long-term precipitation prediction model of the month to be predicted constructed by the construction method for a long-term precipitation prediction model described in any one of the foregoing embodiments, so as to obtain the precipitation amount of the month to be predicted. For specific content, reference can be made to a corresponding part of the foregoing method embodiment, and details are not repeatedly described herein.
[0084] In the long-term precipitation prediction apparatus provided in the embodiments of the present invention, the long-term precipitation prediction model of a corresponding month constructed by using the construction method for the described long-term precipitation prediction model is used to perform precipitation prediction of a month to be predicted. As the constructed long-term precipitation prediction model is constructed by using a false discovery rate control method of a multi-hypothesis test in combination with a random forest model, the prediction accuracy and reliability are relatively high. Hence, the long-term precipitation prediction apparatus has a relatively high prediction accuracy.
[0085] For the functional description of the long-term precipitation prediction apparatus provided by the embodiment of the present invention, reference can be made to the description of the long-term precipitation prediction method in the foregoing embodiment.
[0086] An embodiment of the present invention further provides a storage medium having a computer program 601 stored thereon, as shown in figure 9, and when executing an instruction, the processor implements the steps of the construction method for a long-term precipitation prediction model and a long-term precipitation prediction method in the described embodiments. The storage medium further stores audio and video stream data, feature frame data, interaction request signaling, encrypted data and a pre-set data size, and the like. The storage medium can be a magnetic disk, a compact disc, a read-only memory (ROM), a random access memory (RAM), a flash memory, a hard disk drive (abbreviated as HDD), a solid-state drive (SSD), and the like; and the storage medium can further comprise a combination of the above types of memories [0087] A person skilled in the art can understand that all or a part of the processes of the methods in the embodiments can be implemented by a computer program instructing relevant hardware. The program can be stored in a computer readable storage medium. When the program is implemented, the processes of the methods in the embodiments are performed. The storage medium can be a magnetic disk, a compact disc, a read-only memory (ROM), a random access memory (RAM), a flash memory, a hard disk drive (abbreviated as HDD), a solid-state drive (SSD), and the like; and the storage medium can further comprise a combination of the above types of memories.
[0088] An embodiment of the present invention further provides an electronic device. As shown in Fig. 10, the electronic device can comprise a processor 501 and a processor 502. The processor 501 and the processor 502 therein can be connected by means of a bus or in another manner. Connecting by means of a bus is taken as an example in Fig. 10.
[0089] The processor 501 can be a central processing unit (CPU). The processor 501 can also be a chip such as another general processor, a digital signal processor (DSP), an application specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or another programmable logic device, a discrete gate or transistor logic device, or a discrete hardware component, or a combination of the foregoing chips. [0090] As a non-transitory computer readable storage medium, the processor 502 can be configured to store a non-transitory software program, a non-transitory computer executable program, and a module, such as a corresponding program instruction/module in the embodiments of the present invention. The processor 501 runs the non-transitory software program, instruction and module stored in the processor 502, so as to execute various functional applications and data processing of the processor, that is, to implement the construction method for a long-term precipitation prediction model and the long-term precipitation prediction method prediction in the method embodiments.
[0091] The processor 502 can comprise a program storage area and a data storage area, wherein the program storage area can store an operation apparatus and an application program required by at least one function; and the data storage area can store data created by the processor 501, and the like. In addition, the processor 502 can comprise a high-speed random access memory, and can further comprise a non-transitory memory, for example, at least one magnetic disk storage device, a flash memory device, or another non-transitory solid-state storage device. In some embodiments, processor 502 optionally comprises memory remotely located from processor 501, which can be connected to processor 501 via a network. Examples of such networks comprise, but are not limited to, the Internet, intranets, local area networks, mobile communication networks, and combinations thereof 100921 The one or more modules are stored in the processor 502, and when being executed by the processor 501, the processor executes the construction method for a long-term precipitation prediction model and the long-term precipitation prediction method according to the embodiments shown in Fig. Ito Fig. 6.
[0093] Specific details of the foregoing electronic device can be understood with reference to corresponding related descriptions and effects in the embodiments shown in Fig. I to Fig. 6, which are not described herein again.
Claims (9)
- CLAIMSL A constniction method for a long-term precipitation prediction model, comprising: acquiring a sample set, wherein the sample set comprises explanatory variables and dependent variables, each of the explanatory variables comprises a precipitation amount of each month of a current year and an observation value of each month of the current year for a key climate Factor affecting the precipitation amount of the corresponding month, and each of the dependent variables comprises a precipitation amount of the corresponding month of a next year; screening the explanatory variables on the basis of a false discovery rate control method of a multi-hypothesis test and a random forest model, so as to obtain a prediction factor affecting the precipitation amount of the corresponding month of the next year; and performing random forest modeling according to the prediction factor affecting the precipitation amount of the corresponding month of the next year and the precipitation amount of the corresponding month of the next year, and training the model, so as to obtain a long-term precipitation prediction model for the precipitation amount of the corresponding month; wherein the step of screening the explanatory variables on the basis of a false discovery rate control method of a multi-hypothesis test and a random forest model so as to obtain a prediction factor affecting the precipitation amount of the corresponding month of the next year comprises: constructing a knockoff variable for each of the explanatory variables; scoring the explanatory variables and the knockoff variables according to a random forest model, so as to obtain explanatory variable scores and knockoff variable scores; and screening the interpretation variable scores and the knockoff variable scores on the basis of the false discovery rate control method so as to obtain a prediction factor affecting the precipitation amount of the corresponding month of the next year.
- 2. The construction method for a long-term precipitation prediction model according to claim 1, wherein the step of acquiring a sample set comprises: acquiring historical precipitation data and climate factor data; determining, according to the historical precipitation data and the climate factor data, the precipitation amount of each month and the key climate factor affecting the precipitation amount of the corresponding month, determining, according to a key factor affecting the precipitation amount of the corresponding month, an observation value of each month of the current year for the key climate factor affecting the precipitation amount of the corresponding month; and constructing a sample set according to the precipitation amount of each month of the current year, the observation value of each month of the current year for the key climate factor affecting the precipitation amount of the corresponding month, and the precipitation amount of the corresponding month of the next year.
- 3 The construction method for a long-term precipitation prediction model according to claim 2, wherein after acquiring historical precipitation data and climate factor data, the method further comprises: correcting and quality controlling the historical precipitation data and the climate factor data, determining, according to the historical precipitation data and the climate factor data, the precipitation amount of each month and the key climate factor affecting the precipitation amount of the corresponding month, comprising: screening the climate factor data on the basis of a grey correlation analysis algorithm, so as to obtain the key climate factor affecting the precipitation amount of the corresponding month.
- 4. The construction method for a long-term precipitation prediction model according to claim 1, wherein the step of screening the interpretation variable scores and the knockoff variable scores on the basis of the false discovery rate control method so as to obtain a prediction factor affecting the precipitation amount of the corresponding month of the next year comprises: calculating the explanatory variable scores and the knockoff variable scores, so as to obtain a competitive scoring of each of the explanatory variables; and screening the competitive scoring on the basis of a pre-set false rate level, so as to obtain a prediction factor affecting the precipitation amount of the corresponding month of the year.
- 5. A long-term precipitation prediction method, comprising: acquiring a precipitation amount of each month of a previous year of the month to be predicted and an observation value of each month of the previous year for a key climate factor affecting the precipitation amount of the month to be predicted, and inputting the precipitation amount of each month of a previous year and the observation value of each month of the previous year for the key climate factor affecting the precipitation amount of the month to be predicted into the long-term precipitation prediction model of the month to be predicted constructed by the construction method for a long-term precipitation prediction model according to any one of claims 1-4, so as to obtain the precipitation amount of the month to be predicted.
- 6. A construction apparatus for a long-term precipitation prediction model, comprising: a sample acquisition module, wherein the sample acquisition module is used for acquiring a sample set, the sample set comprises explanatory variables and dependent variables, each of the explanatory variables comprises a precipitation amount of each month of a current year and an observation value of each month of the current year for a key climate factor affecting the precipitation amount of the corresponding month, and each of the dependent variables comprises a precipitation amount of the corresponding month of a next year; a screening module, wherein the screening module is used for screening the explanatory variables on the basis of a false discovery rate control method of a multi-hypothesis test and a random forest model, so as to obtain a prediction factor affecting the precipitation amount of the corresponding month of the next year, and a modeling module, wherein the modeling module is used for performing random forest modeling according to the prediction factor affecting the precipitation amount of the corresponding month of the next year and the precipitation amount of the corresponding month of the next year, and training the model, so as to obtain a long-term precipitation prediction model for the precipitation amount of the corresponding month.
- 7. A long-term precipitation prediction apparatus, comprising: a data acquisition module, wherein the data acquisition module is used for acquiring a precipitation amount of each month of a previous year of the month to be predicted and an observation value of each month of the previous year for a key climate factor affecting the precipitation amount of the month to be predicted, and a prediction module, wherein the prediction module is used for inputting the precipitation amount of each month of a previous year and the observation value of each month of the previous year for the key climate factor affecting the precipitation amount of the month to be predicted into the long-term precipitation prediction model of the month to be predicted constructed by the construction method for a long-term precipitation prediction model according to any one of claims 1-4, so as to obtain the precipitation amount of the month to be predicted.
- 8. A computer readable storage medium having a computer Instruction stored thereon, wherein the computer instruction is used to enable a computer to execute the construction method for a long-term precipitation prediction model according to any one of claims 1-4 and the long-term precipitation prediction method according to claim 5.
- 9. An electronic device, comprising a memory and a processor, wherein the memory having a computer instruction stored thereon and the processor are in communication connection with each other; and when executing the computer instruction, the processor executes the construction method for a long-term precipitation prediction model according to any one of claims 1-4 and the long-term precipitation prediction method according to claim S.
Applications Claiming Priority (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202210110058.6A CN114118640B (en) | 2022-01-29 | 2022-01-29 | Long-term precipitation prediction model construction method, long-term precipitation prediction method and device |
PCT/CN2022/112632 WO2023284887A1 (en) | 2022-01-29 | 2022-08-16 | Long term precipitation prediction model establishing method, and long-term precipitation prediction method and apparatus |
Publications (2)
Publication Number | Publication Date |
---|---|
GB202302648D0 GB202302648D0 (en) | 2023-04-12 |
GB2615205A true GB2615205A (en) | 2023-08-02 |
Family
ID=80361714
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
GB2302648.7A Pending GB2615205A (en) | 2022-01-29 | 2022-08-16 | Long term precipitation prediction model establishing method, and long-term precipitation prediction method and apparatus |
Country Status (3)
Country | Link |
---|---|
CN (1) | CN114118640B (en) |
GB (1) | GB2615205A (en) |
WO (1) | WO2023284887A1 (en) |
Families Citing this family (11)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN114118640B (en) * | 2022-01-29 | 2022-04-26 | 中国长江三峡集团有限公司 | Long-term precipitation prediction model construction method, long-term precipitation prediction method and device |
CN116128099B (en) * | 2022-11-28 | 2023-09-12 | 南京信息工程大学 | Artificial intelligence-based China short-term climate prediction method and system |
CN116485010B (en) * | 2023-03-20 | 2024-04-16 | 四川省雅安市气象局 | S2S precipitation prediction method based on cyclic neural network |
CN116449460B (en) * | 2023-04-23 | 2023-10-10 | 水利部交通运输部国家能源局南京水利科学研究院 | Regional month precipitation prediction method and system based on convolution UNet and transfer learning |
CN116313131B (en) * | 2023-05-24 | 2023-09-15 | 山东大学 | Brain network difference recognition system, equipment and storage medium based on imitation variables |
CN116611588B (en) * | 2023-07-19 | 2023-09-22 | 水利部水利水电规划设计总院 | Precipitation multi-driving factor segmentation rating optimization forecasting method and system |
CN117034780B (en) * | 2023-08-31 | 2024-04-19 | 江苏省气候中心 | Multi-scale sub-season precipitation prediction method based on deep learning |
CN117314184B (en) * | 2023-09-19 | 2024-04-30 | 沈阳农业大学 | Method, device and equipment for determining time-space change result of annual total primary productivity |
CN117542443B (en) * | 2023-09-27 | 2024-05-24 | 中国农业大学 | Method and device for balancing yield and relieving nitrogen pollution and electronic equipment |
CN117708113B (en) * | 2024-02-06 | 2024-05-17 | 中国电建集团西北勘测设计研究院有限公司 | Precipitation data construction method |
CN117933299A (en) * | 2024-03-19 | 2024-04-26 | 南京大学 | Optimal climate mode artificial intelligence identification method and system for climate prediction |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20130231906A1 (en) * | 2010-10-04 | 2013-09-05 | Ofs Fitel, Llc | Statistical Prediction Functions For Natural Chaotic Systems And Computer Models Thereof |
CN109376913A (en) * | 2018-09-30 | 2019-02-22 | 北京市天元网络技术股份有限公司 | The prediction technique and device of precipitation |
CN112884209A (en) * | 2021-01-29 | 2021-06-01 | 河海大学 | Weather method and mathematical statistics method-based medium and long-term rainfall forecasting method |
CN113537600A (en) * | 2021-07-20 | 2021-10-22 | 浙江省水利水电勘测设计院 | Medium-and-long-term rainfall forecast modeling method based on whole-process coupled machine learning |
CN114118640A (en) * | 2022-01-29 | 2022-03-01 | 中国长江三峡集团有限公司 | Long-term precipitation prediction model construction method, long-term precipitation prediction method and device |
Family Cites Families (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN105044361B (en) * | 2015-08-14 | 2017-07-28 | 山东省肿瘤防治研究院 | A kind of diagnostic marker and its screening technique for being suitable for esophageal squamous cell carcinoma early diagnosis |
CN112217787B (en) * | 2020-08-31 | 2022-11-04 | 北京工业大学 | Method and system for generating mock domain name training data based on ED-GAN |
-
2022
- 2022-01-29 CN CN202210110058.6A patent/CN114118640B/en active Active
- 2022-08-16 GB GB2302648.7A patent/GB2615205A/en active Pending
- 2022-08-16 WO PCT/CN2022/112632 patent/WO2023284887A1/en unknown
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20130231906A1 (en) * | 2010-10-04 | 2013-09-05 | Ofs Fitel, Llc | Statistical Prediction Functions For Natural Chaotic Systems And Computer Models Thereof |
CN109376913A (en) * | 2018-09-30 | 2019-02-22 | 北京市天元网络技术股份有限公司 | The prediction technique and device of precipitation |
CN112884209A (en) * | 2021-01-29 | 2021-06-01 | 河海大学 | Weather method and mathematical statistics method-based medium and long-term rainfall forecasting method |
CN113537600A (en) * | 2021-07-20 | 2021-10-22 | 浙江省水利水电勘测设计院 | Medium-and-long-term rainfall forecast modeling method based on whole-process coupled machine learning |
CN114118640A (en) * | 2022-01-29 | 2022-03-01 | 中国长江三峡集团有限公司 | Long-term precipitation prediction model construction method, long-term precipitation prediction method and device |
Non-Patent Citations (1)
Title |
---|
The randomForest Package [online]. Liaw, Andy. 2008 [retrieved on 2023-03-31]. Retrieved from <https://web.archive.org/web/20090206000200/http://cran.r-project.org/web/packages/randomForest/randomForest.pdf> * |
Also Published As
Publication number | Publication date |
---|---|
GB202302648D0 (en) | 2023-04-12 |
WO2023284887A1 (en) | 2023-01-19 |
CN114118640B (en) | 2022-04-26 |
CN114118640A (en) | 2022-03-01 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
GB2615205A (en) | Long term precipitation prediction model establishing method, and long-term precipitation prediction method and apparatus | |
Yan et al. | Corrosion rate prediction and influencing factors evaluation of low-alloy steels in marine atmosphere using machine learning approach | |
CN111382929A (en) | Method and device for constructing river diatom bloom early warning model | |
Monteleoni et al. | Climate informatics: accelerating discovering in climate science with machine learning | |
WO2021179742A1 (en) | Ozone missing data interpolation method, apparatus and device | |
CN114492211B (en) | Residual oil distribution prediction method based on autoregressive network model | |
CN110969600A (en) | Product defect detection method and device, electronic equipment and storage medium | |
Cheng et al. | Discrete principal‐monotonicity inference for hydro‐system analysis under irregular nonlinearities, data uncertainties, and multivariate dependencies. Part I: methodology development | |
CN114911788B (en) | Data interpolation method and device and storage medium | |
CN112180471A (en) | Weather forecasting method, device, equipment and storage medium | |
CN116383773A (en) | Data center energy efficiency anomaly detection method, system and medium based on self-adaptive prediction interval | |
Peña et al. | Estimation of analysis and forecast error variances | |
CN115759365A (en) | Photovoltaic power generation power prediction method and related equipment | |
JP5110891B2 (en) | Statistical prediction method and apparatus for influent water quality in water treatment facilities | |
Krock et al. | Nonstationary seasonal model for daily mean temperature distribution bridging bulk and tails | |
CN115980755A (en) | Storm approach extrapolation method and device based on ConvLSTM | |
CN114529035A (en) | CART-based wind speed forecasting method of multi-mode integrated model | |
CN114169646B (en) | Water bloom prediction method, device, electronic equipment and computer readable storage medium | |
CN114611799B (en) | Time sequence neural network new energy output multi-step prediction method based on supervised learning | |
US20220243347A1 (en) | Determination method and determination apparatus for conversion efficiency of hydrogen production by wind-solar hybrid electrolysis of water | |
WO2022222230A1 (en) | Indicator prediction method and apparatus based on machine learning, and device and storage medium | |
CN115115107A (en) | Photovoltaic power prediction method and device and computer equipment | |
CN114154697A (en) | House maintenance resource prediction method and device, computer equipment and storage medium | |
Wang et al. | A feasibility study for the construction of an atmospheric precipitable water vapor model based on the neural network technology | |
van Proosdij et al. | Unequal contribution of widespread and narrow-ranged species to botanical diversity patterns |