CN109902885A - Typhoon prediction method based on deep learning mixed CNN-LSTM model - Google Patents

Abstract

The invention discloses a typhoon prediction method based on a deep learning hybrid CNN-LSTM model, which aims to solve the problems of unstable prediction results and low accuracy of the existing typhoon prediction methods. The technical solution is to first obtain the historical tropical cyclone optimal path data set and the global atmospheric and oceanic variable data to form the data set of the training model, then construct the typhoon prediction model of the hybrid CNN-LSTM, and then use the data set of the training model to analyze the hybrid CNN-LSTM. The LSTM typhoon prediction model is trained and fitted, and finally the trained hybrid CNN‑LSTM typhoon prediction model is used to predict whether a typhoon will form and its path and intensity after formation. By adopting the present invention, the existing public data set and deep learning framework can be used to construct and predict the model conveniently, and the present invention can directly use the typhoon data and atmospheric and oceanic variable data in recent decades to train the constructed model. Use the trained model to effectively improve the accuracy of typhoon prediction.

Description

Typhoon forecast method based on deep learning mixing CNN-LSTM model

Technical field

The present invention relates to one kind based on deep learning mixing convolutional neural networks (CNN) and shot and long term Memory Neural Networks (LSTM) the typhoon forecast method of model belongs to computer and meteorological, ocean crossing domain.

Background technique

Typhoon not only will affect offshore activities as a kind of Extreme Weather Events, but also can give the life of maritime people Living and urban economy causes heavy losses.Therefore, typhoon research and prediction are always the emphasis of each coastal state's concern.Typhoon Prediction is usually to predict the intensity of typhoon and path, and in recent years, path prediction achieves bigger progress, and typhoon Prediction of strength is but still without too big improvement.

In order to provide the early warning of typhoon disaster, accurate typhoon formation, intensity, path prediction technique are always meteorology The focal point in field.Due to the complexity and variability of typhoon system, existing typhoon forecast method is generally difficult to typhoon Accomplish accurately to predict.Existing typhoon forecast method is generally divided into three classes, and one kind is the prediction technique based on numerical model, such as base In region mode Hurricane Weather and Research Forecasting Model (HWRF), Global Model ECMWF global model (EMX), intergration model (GFS) prediction technique etc.；One kind is the prediction technique based on statistical analysis, Such as weather persistence model Climatology and Persistence model (CILIEPER5), statistics dynamic model Statistical Hurricane Intensity Prediction Scheme (SHIPS) etc.；There are also one kind to be based on engineering The prediction technique of habit, such as logistic regression Logistic Regression (LR) and artificial neural network Artificial Neural Network (ANN) etc..Gone out based on the prediction technique of numerical model from the physical background field of atmosphere Hair, the activity to be detected a typhoon by solving hydrodynamic equations.And it is existing pre- with machine learning based on statisticalling analyze Survey method is usually some significant variables found out and influenced during Typhoon Activity, then excavate these variables and typhoon formation it Between correlation, with to after typhoon formation and formation path and intensity predict.Such as document " Wijnands J S, Qian G, Kuleshov Y.Variable Selection for Tropical Cyclogenesis Predictive Modeling [J] .Monthly Weather Review, 2016,144. " (the i.e. variable choosings of tropical cyclone generation prediction modeling Select, " the monthly weather review " page 144 published in 2016) it describes in a text and a kind of generates predictive variable for tropical cyclone Preference pattern, number is analyzed by using history tropical cyclone optimal path data set and European medium-range numerical forecast center again According to when qualified tropical cyclone is selected from history tropical cyclone optimal path data set first generating and be mobile Between and position, then corresponding environment field variable is matched from analyzing again according to time and position in data, then with probability artwork After type screens variable, machine learning model is established to predict the shape of tropical cyclone with LR using the variable after screening At.Prediction for path and intensity, now more commonly used method is Statistical-dynamic method, and this method is with Numerical Weather mould Formula is to rely on, and the environment field parameter that use pattern operation obtains establishes Statistical Prediction Model.It is most represented in typhoon operational forecast Property be SHIPS model, what which selected when initially proposing is the duration variable of several Climatologicals, such as acts the heat at quarter of giving the correct time Julian date, tropical cyclone when with cyclone intensity, longitude and latitude, moving direction, generation shift to vector and other dynamical variables, Such as the opposite whirlpool of vertical wind shear, 200 hPa shell of compression between maximum possible intensity, 850~20,000 pa (hPa) shell of compression Momentum spread etc. establishes statistical regression model using these variables come the path to detect a typhoon and intensity.The statistics of the model is returned Return coefficient that will be updated with new sample in annual typhoon season, and can constantly be improved in operational forecast.

Although existing typhoon forecast method has considerable effect for current typhoon forecast, but still exists very More shortcomings.The difference of numerical model method parameter scheme and primary condition can have very big shadow to the accuracy of prediction It rings, so that there are unstability for prediction result.Existing statistical method is difficult to consider the non-linear relation between each variable, machine Learning method does not consider the temporal correlation of each variable, so that the accuracy of prediction decreases.Therefore these types of method is equal It is difficult to obtain the higher prediction effect of more stable and accuracy.In addition, existing method substantially focuses on that typhoon influence becomes The selection of amount, does not there is the improvement of relatively deep to model, and too simple model is likely difficult to portray the complexity of Typhoon Activity Process.Therefore how to provide one efficient, easy to carry out and that accuracy is high typhoon forecast method is those skilled in the art pole The technical issues of to pay close attention to.

Summary of the invention

The technical problem to be solved by the present invention is to for existing typhoon forecast method prediction result, there are unstability, standards Exactness problem to be improved proposes a kind of typhoon forecast method based on deep learning mixing CNN-LSTM model.This method The Atmosphere and Ocean variable data of usage history tropical cyclone optimal path data set and the whole world, by using 3DCNN (three-dimensional CNN) The space characteristics for extracting atmospheric variable data extract the space characteristics of ocean variable data with 2DCNN (two-dimentional CNN), use LSTM The time serial message during typhoon occurrence and development is extracted, the typhoon forecast mould of a mixing CNN-LSTM is thus constructed Type, and path after whether typhoon is formed and formed using the model and intensity are predicted.

Technical solution is: the Atmosphere and Ocean variable number of acquisition history tropical cyclone optimal path data set and the whole world first According to, form the data set of training pattern, then building mixing CNN-LSTM typhoon forecast model, recycle the number of training pattern It is trained fitting according to typhoon forecast model of the collection to mixing CNN-LSTM, finally using the platform of trained mixing CNN-LSTM Wind prediction model whether typhoon is formed and is formed after path and intensity predict.

Specific technical solution of the present invention the following steps are included:

The first step obtains tropical cyclone optimal path data set, selects tropical cyclone optimal path record, matching record pair The Atmosphere and Ocean variable data answered, data set needed for constructing training pattern, method is:

1.1 download history tropical cyclone optimal path data set in recent years from tropical cyclone official website.The selection torrid zone The tropical cyclone optimal path data set in the region that sea area is divided is generated in cyclone optimal path data set by tropical cyclone (element is tropical cyclone optimal path record).Including different year in each region tropical cyclone optimal path data set Tropical cyclone optimal path record, each year include the tropical cyclone optimal path record of different item numbers again.Single tropical cyclone Optimal path record is containing the tropical cyclone from being formed the All Time of extinction, longitude and latitude position and strength information.

The 1.2 a certain fixed area for enabling needs detect a typhoon are region A₁, A₁Tropical cyclone optimal path data set be Q, It checks whether the maximum intensity in Q on single tropical cyclone optimal path reaches 64kt (32.7m/s) one by one, maximum intensity is reached Single tropical cyclone record (i.e. typhoon record) to 64kt is put to typhoon data set Q₁, maximum intensity is reached into 34kt The tropical cyclone single record (i.e. tropical storm record) that (17.2m/s) is not up to 64kt is put to tropical storm data set Q₂, by This constructs typhoon data set Q₁, tropical storm data set Q₂, remaining record for being unsatisfactory for condition casts out.

1.3 downloading Q₁Middle whole typhoon records the time of all moment record points, longitude and latitude position, is become using Atmosphere and Ocean Data set generation method is measured according to Q₁Middle whole typhoon records the time of all moment record points, center longitude position carries out group It knits, Q needed for obtaining model training₁Corresponding Atmosphere and Ocean variable data collectionMethod is:

1.3.1 by Q₁In typhoon record carry out it is randomly ordered, enable N₁For Q₁Typhoon record number, initialization typhoon note Record serial number n_p=1；

1.3.2 enablingIt is n-th_pThe record point number of typhoon record, initialization n-th_pThe record point sequence of typhoon record Number

1.3.3 according to n-th_pThe of article typhoon recordThe time of a record pointLongitude station Position LatitudeUsing single-point variable data method for down loading, from weather forecast center website, (such as: European medium-range weather is pre- Report center) the corresponding Atmosphere and Ocean variable data of selection downloading, single-point variable data method for down loading is:

1.3.3.1 the range of Atmosphere and Ocean variable data downloading is set as with n-th_pThe of article typhoon recordA record point Longitude and latitude position centered on w ° × w ° longitude and latitude grid, w is positive integer less than 180, the longitude of w ° of expression data cover Span or latitude span.

1.3.3.2 the resolution ratio of Atmosphere and Ocean variable data downloading is set as n-th_pThe of article typhoon recordA record point U ° × u ° of the highest resolution that can be got.The u ° of longitude interval or latitude interval indicated between lattice point, region A₁Interior lattice point Number is

1.3.3.3 the atmospheric variable of downloading is set as n-th_pThe of article typhoon recordThe variable of a record, is denoted as p₁, p₂..., p_i..., p_I, I is atmospheric variable species number, and I is positive integer, and selected variable is relevant to typhoon occurrence and development Atmospheric variable；Layer is n-th where setting the atmospheric variable space of downloading_pThe of article typhoon recordA record it is retrievable l₁, l₂..., l_i..., l_zLayer, z are the atmospheric variable number of plies, and z is positive integer.

1.3.3.4 the ocean variable of downloading is set as n-th_pThe of article typhoon recordThe variable of a record, is denoted as s₁, s₂..., s_i..., s_J, J is ocean variable species number, and J is positive integer, and selected variable is relevant to typhoon occurrence and development Ocean variable；Layer where setting the ocean variable space of downloading is n-th_pThe of article typhoon recordA record it is retrievable f₁, f₂..., f_i... f_WLayer, W are the ocean variable number of plies, and W is positive integer.

1.3.3.5 n-th is downloaded_pThe of article typhoon recordThe corresponding atmospheric variable data X of a record point^PAnd ocean Variable data X^S, by X^PAnd X^SIt is packaged composition n-th_pThe of article typhoon recordThe time of a record point and longitude and latitude are big accordingly Qihai ocean variable data, is denoted as It is n-th_pThe of article typhoon recordA record The time of point,It is n-th_pThe of article typhoon recordThe longitude of a record point,It is n-th_pTyphoon note The of recordThe latitude of a record point.

1.3.4 enabling

If 1.3.5 n_p≤N₁AndTurn 1.3.3；If n_p≤N₁ButEnable n_p=n_p+ 1, it goes to step 1.3.2；If n_p> N₁, execute 1.3.6.

1.3.6 will Packing group As a data set, as Q₁Corresponding Atmosphere and Ocean variable data collectionContaining N₁All record points pair of typhoon The Atmosphere and Ocean variable data answered.

1.4 downloading Q₂Whole tropical storms record the time of all moment record points, center longitude position, using step The 1.3 Atmosphere and Ocean variable data set creation methods, according to Q₂Whole tropical storms record all moment record point time, Longitude and latitude position carries out tissue, Q needed for obtaining model training₂Corresponding Atmosphere and Ocean variable data collection

1.5 according to Q₁, Q₂AndFor the data of the different different prediction models of prediction target building training Collection, method is:

If 1.5.1 prediction target is whether typhoon forms, data needed for whether training typhoon forms prediction model are constructed Collect R₁, method is:

1.5.1.1 according to Q₁AndConstruct the required data set R of training pattern₁₁, method is:

1.5.1.1.1 typhoon record number n is initialized_p=1；Initialize R₁₁Input data serial number q=1；

1.5.1.1.2 from Q₁In select n-th_pTyphoon record；FromIn select n-th_pTyphoon records corresponding atmosphere Ocean variable data It is packaged, is recorded as

1.5.1.1.3 n-th is enabled_pThe numbered positions of record point of first intensity greater than 64kt are in typhoon record

1.5.1.1.4 it selectes predicted time and walks k, look-back time walks b.Enabling k=T/6, T is predicted time, and for 6 it is just whole Several times, any positive integer of the b between 1~k.

1.5.1.1.5 ifFromIn pick out n-th_pThe in article record The corresponding Atmosphere and Ocean variable data of a record point is packaged, composition n-th_pTyphoon record For input data constructed by training pattern, it is recorded as X '_q, X '_qIndicate R₁₁In q-th of input data, enable q=q+1, turn 1.5.1.1.6；IfDirectly turn 1.5.1.1.6.

1.5.1.1.6 n is enabled_p=n_p+ 1, if n_p≤ N1, turns 1.5.1.1.2；If n_p> N₁, illustrate to have obtainedM₁For R₁₁In input data number, M₁≤N₁, turn 1.5.1.1.7.

1.5.1.1.7 willIt is packaged, is utilized Q₁In all typhoon recording needles to training pattern institute The input data set of building.By M₁A input data tagged 1, is utilized Q₁For label constructed by training pattern Data set.Input data set and label data collection have collectively constituted data set R needed for whether trained typhoon forms model₁₁, R₁₁ Containing M₁A input data and M₁A label data.

1.5.1.2 according to Q₂AndConstruct the required data set R of training pattern₁₂, method is:

1.5.1.2.1 N is enabled₂For Q₂Tropical storm record number, initialize tropical storm record number m_p=1；Initialization R₁₂Input data serial number h=1；

1.5.1.2.2 from Q₂In select m_pTropical storm record, enables m_pThe number of tropical storm record record point ForFromIn select m_pTropical storm records corresponding Atmosphere and Ocean variable dataIt is packaged, note For

1.5.1.2.3 enabling m_pThe numbered positions of the record point of first maximum of intensity are in tropical storm record

1.5.1.2.4 it selectes predicted time and walks k, look-back time walks b.Enable k be T/6, T is predicted time, and for 6 it is just whole Several times, any positive integer of the b between 1~k.

1.5.1.2.5 ifFromIn pick out m_pThe in article record The corresponding Atmosphere and Ocean variable data of a record point is packaged, and forms m_pTropical storm Recording needle is recorded as X ' to input data constructed by training pattern_h, X '_hIndicate R₁₂In h-th of input data, enable h=h+1, Turn 1.5.1.2.6；IfDirectly turn 1.5.1.2.6.

1.5.1.2.6 m is enabled_p=m_p+ 1, if m_p≤N₂, turn 1.5.1.2.2；If m_p> N₂, illustrate to have obtainedM₂For R₁₂In input data number, M₂≤N₂, turn 1.5.1.2.7.

1.5.1.2.7 willIt is packaged, is utilized Q₂In all tropical storm recording needles to training mould Input data set constructed by type.By M₂A input data tagged 0, is utilized Q₂For constructed by training pattern Label data collection.Input data set and label data collection have collectively constituted data set needed for whether trained typhoon forms model R₁₂, R₁₂Containing M₂A input data and M₂A label data.

1.5.1.3 by R₁₁And R₁₂Merge into data set R needed for whether trained typhoon forms prediction model₁, R₁Containing model Required I₁A input data, I₁A label data, I₁=M₁+M₂；Turn 1.6.

If 1.5.2 prediction target is typhoon track, data set R needed for constructing training typhoon track prediction model₂, method It is: to Q₁In all typhoons record latitude and longitude value and corresponding Atmosphere and Ocean variable dataDislocation processing is carried out, at dislocation The method of reason is:

1.5.2.1 typhoon record number n is initialized_p=1；Enable I₂Indicate R₂In input data number, initialization data set R₂The serial number i of middle input data_p=1；

1.5.2.2 from Q₁In at random select n-th_pTyphoon record.

1.5.2.3 n-th is enabled_pThe numbered positions recorded a little in typhoon record areInitializationThis is enabled to record Above the total number of all record points is

1.5.2.4 it selectes predicted time and walks k, look-back time walks b.K is T/6, and the positive integer times that T is 6, b is between 1~k Any positive integer.

1.5.2.5 ifFromA record point starts, selection The Atmosphere and Ocean variable data of total b+1 record point, which is packaged, is used as R₂In i-th_pA input data, is recorded asSelection theThe typhoon longitude and latitude position of point is recorded as R₂In i-th_pThe label data of a input data, note Record isTurn 1.5.2.6；IfButSelectionAltogether The Atmosphere and Ocean variable data of b+1 record point is as R₂In i-th_pThe input data of a sample, is recorded asIt willThe typhoon longitude and latitude position value of a record point is assigned to 0, and as R₂In i-th_pA input number According to label data, be recorded asTurn 1.5.2.6；IfEnable n_p=n_p+ 1, turn 1.5.2.7.

1.5.2.6 enablingi_p=i_p+ 1, turn 1.5.2.5.

If 1.5.2.7 n_p≤ N1, turns 1.5.2.2；If n_p> N₁, turn 1.5.2.8.

1.5.2.8 willIt is packaged, is utilized Q₁In all typhoon recording needles to training pattern institute structure The path input data set built.It willIt is packaged, is utilized Q₁In all typhoon recording needles to training pattern institute The path label data set of building.Input data set and label data collection have collectively constituted needed for trained typhoon track prediction model Data set R₂, R₂Containing I₂A path input dataAnd I₂A path label dataPath label data are latitude and longitude value；Turn 1.6.

If 1.5.3 prediction target is intensity of typhoon, data set R needed for constructing training intensity of typhoon prediction model₃, method It is that the dislocation processing method is walked using 1.5.2 to Q₁In all typhoons record intensity value and corresponding Atmosphere and Ocean variable number According toDislocation processing is carried out, intensity input data set and intensity label data needed for training intensity of typhoon prediction model are obtained Collection, constitutes data set R by intensity input data set and intensity label data set₃, intensity input data set is containing I₂A intensity inputs number According toIntensity label data set contains I₂A intensity label dataIntensity label data is intensity value；Turn 1.6.

1.6 couples of R₁, R₂, R₃It is pre-processed；Respectively to R₁Input data set, R₂Input data set and R₂Number of tags According to collection, R₃Input data set and R₃Label data collection be normalized by formula (1),

Wherein R indicates R₁, R₂, R₃In any one data set input data set or R₂, R₃In any one data set Label data collection, R_maxIndicate the maximum value of the data set, R_minIndicate the minimum value of the data set.R₁, R₂, R₃It is pretreated Data set is respectively R '₁, R '₂, R '₃；

Second step, building deep learning mix CNN-LSTM model, and method is:

2.1 build deep learning environment.Deep learning library Keras based on Python is installed on the server, and is used Google deep learning frame TensorFlow is as rear end.

The 2.2 deep learning library Keras building deep learnings based on Python mix CNN-LSTM typhoon forecast model, side Method is:

2.2.1 learn the 3DCNN module of atmospheric variable space characteristics in building mixing CNN-LSTM model:

The input layer for being 1 layer of Keras by the configuration settings of neural network in 3DCNN module, m_ALayer Three dimensional convolution layer (Conv3d layers), m_ALayer batch regularization layer (BatchNormalization layers), 1 layer of tiling layer (flatten layers), 1 layer connects entirely Connect layer (Dense layers).The 1st layer of Three dimensional convolution layer is set to contain3 a dimensions convolution kernel (i.e. filter), the size of convolution kernel areStep-length isThe 2nd layer of Three dimensional convolution layer is set to contain3 a dimension convolution kernels (filter Device), the size of convolution kernel isStep-length isSet m_ALayer Three dimensional convolution layer contains3 a dimensions convolution kernel (i.e. filter), the size of convolution kernel areStep-length isOne layer batch of regularization layer is added after every layer of Three dimensional convolution layer, and all layers of neuron selection is sharp Function living is ' Relu ' function, and concrete form is Relu (x)=max (0, x) (taking the maximum value in 0 and x).' Relu ' function On the one hand the non-linear property that model can be enhanced, on the other hand can be enhanced the sparsity of network.Set the nerve of full articulamentum First number is d_AIt is a, wherein parameter It is positive integer.

2.2.2 learn the 2DCNN module of ocean variable space feature in building mixing CNN-LSTM model:

The input layer for being 1 layer of Keras by the configuration settings of neural network in 2DCNN module, m_BLayer two-dimensional convolution layer (Conv2d layers), m_BLayer batch regularization (BatchNormalization layers), 1 layer of tiling layer (flatten layers) and 1 layer of full connection (Dense layers) composition of layer, as shown in the module for handling ocean variable in Fig. 3.The 1st layer of two-dimensional convolution layer is set to containA 2D The size of convolution kernel, convolution kernel isStep-length isThe 2nd layer of two-dimensional convolution layer is set to containA 2D convolution The size of core, convolution kernel isStep-length isEvery layer of two dimension volume ..., setting m_BLayer two-dimensional convolution layer containsA 2D convolution kernel, the size of convolution kernel areStep-length isAdd after every layer of two-dimensional convolution layer One layer of batch regularization layer, and the activation primitive that all layers of neuron selects is ' Relu ' function.Set the nerve of full articulamentum First number is d_BIt is a, wherein

It is positive integer.

2.2.3 learn the LSTM module of temporal aspect in building mixing CNN-LSTM model:

The structure of neural network in LSTM module is set as the LSTM of 1 layer of Keras layer.The neuron number that LSTM is arranged is d_CA, the activation primitive that all layers of neuron selects is still ' Relu ' function.

2.2.4 by 3DCNN module, 2DCNN module, LSTM module passes through the Sequential sequent modular built in Keras Connection, and one layer of output layer is added, set output layer neuron number as 1, be mapped as single value to export, neuron selects The activation primitive selected is ' sigmoid ' function σ (x), and concrete form is(truth of a matter that e is natural logrithm), Model after connection is to mix CNN-LSTM model.Wherein, the input of 3DCNN module is R '₁, R '₂, R '₃Middle input data set Atmospheric variable data, export as airspace feature vector O^P, airspace feature vector O^PVector length is full articulamentum Neuron number d_A.The input of 2DCNN module is R '₁, R '₂, R '₃The ocean variable data of middle input data set, exports as atmosphere Spatial signature vectors O^S, airspace feature vector O^SVector length is the neuron number d of full articulamentum_B.LSTM module it is defeated Enter the merging vector O=[O for 3DCNN module and 2DCNN module output vector^P, O^S], merging vector O is Atmosphere and Ocean space Feature vector, O are got by (Merge layers) merging of merging layer of Keras, and the length of O is d_A+d_B.The output of LSTM module is 1 A output mapping value.

Third step, be fitted to obtain by constructed data set training mixing CNN-LSTM model typhoon whether formed it is pre- Model, typhoon tracks forecast model, intensity of typhoon prediction model are surveyed, method is:

3.1 according to demand by R '₁, R '₂, R '₃In data upset at random after (input data and label data need one a pair of Answer), the ratio cut partition by a: b is training set R '₁₁, R '₂₁, R '₃₁With test set R '₁₂, R '₂₂, R '₃₂, wherein a, b be positive integer and a≥b。R′₁₁With R '₁₂Contain N respectively_FAnd N_F' a input data and label data, R '₂₁With R '₂₂Contain N respectively_LAnd N_L' a input number According to and label data, R '₃₁With R '₃₂Equally contain N respectively_LAnd N_L' a input data and label data.

3.2 utilize R '₁Training mixing CNN-LSTM model obtains whether typhoon forms prediction model to fitting, and method is such as Under:

3.2.1 it selects to intersect loss function of the entropy function as mixing CNN-LSTM model, functional form is as follows:

Wherein, loss function value is loss_value1, N_FFor R '₁₁Input data or label data total number, i_FFor R′₁₁The serial number of middle data, 1≤i_F≤N_F,For R '₁₁In i-th_FA true label data,For the number of tags of model prediction According to,Expression pairTake logarithm.

3.2.2 evaluation index of the accuracy rate (ACC) as model quality is selected:

Wherein, N_AFor R '₁₂The middle sum by the correct label data of model prediction, N_F' it is R '₁₂Label data sum.

3.2.3 by R '₁Input mixing CNN-LSTM model, is trained mixing CNN-LSTM model, fitting obtains platform The no formation prediction model of wind, the loss function of model are to intersect entropy function, evaluation index ACC.

It comprises the concrete steps that:

3.2.3.1 loop initialization number serial number nc=1；

3.2.3.2 initializing the number of iterations epoch=10；Epoch value interval is 10~k_e, and be positive integer；k_eIt is all Positive integer, and k_e≥10；

3.2.3.3 initialization batch size batchsize=32；The value interval of batchsize is 32~N_F, and be 32 Positive integer times；

3.2.3.4 initializing learning rate lr=0.1；The value interval of lr isk_lIt is just whole between 1~9 Number.

3.2.3.5 model.compile (loss, optimizer (lr), the metrics) function built in Keras is called Compiling mixing CNN-LSTM model, the parameter setting of function are as follows: loss function loss is to intersect entropy function, optimization algorithm Optimizer is Adam algorithm, and learning rate lr is lr, and evaluation index metrics is ACC.

3.2.3.6 by call Keras built in model.fit (data_train, labels_train, epochs, Batch_size) function mixes CNN-LSTM model to be fitted, and obtains the loss function value loss_valuel of training set.By letter Several parameter setting are as follows: training data data_train is R '₁₁N_FA input data, training label data labels_train For R '₁₁N_FA label data, the number of iterations epochs are epoch, and crowd size batch_size is batchsize.

3.2.3.7 by calling model.evaluate (data_test, the labels_test) function built in Keras Assessment mixing CNN-LSTM model, obtains the ACC of mixing CNN-LSTM model.By the parameter setting of function are as follows: test data Data_test is R '₁₂N_F' a input data, test data labels_test are R '₁₂N_F' a label data.

If 3.2.3.8 epoch≤k_eAnd batchsize≤N_FAndRecord n-th_cThe epoch of secondary circulation, Batchsize, lr, the loss function value loss_valuel of training set and the ACC of test set are called built in Keras Model.save (filename) function saves n_cSecondary to recycle obtained model, filename is set as model name.Enable n_c=n_c+ 1, lr=lr × 0.1, turn 3.2.3.5；If epoch≤k_eAnd batchsize≤N_FButEnable batchsize= Batchsize+32 turns 3.2.3.4；If epoch≤k_eBut batchsize > N_F, epoch=epoch+1 is enabled, 3.2.3.2 is turned； If epoch > k_e, turn 3.2.3.9.

3.2.3.9 by the 1st, 2 ..., n_c..., epoch, batchsize, lr, loss_ that C circulation is recorded Value1, ACC are integrated into the complete documentation of a model of fit.Searched out from complete documentation the smallest loss_value1 and Cycle-index serial number n ' corresponding to maximum ACC_c.By the n-th '_cThe epoch of secondary preservation, batchsize, lr is as optimal ginseng Number, by the n-th '_cWhether the model of secondary preservation forms prediction model as typhoon.

3.3 utilize R '₂Training mixing CNN-LSTM model, fitting obtain typhoon tracks forecast model, the method is as follows:

3.3.1 select root-mean-square error function as the loss function of typhoon tracks forecast model:

Wherein, loss function value is loss_value2, N_LFor R '₂₁Input data or label data total number, i_LFor R′₂₁The serial number of middle data, 1≤i_L≤N_L, Y_iLFor R '₂₁In i-th_LA true label data,It is i-th_LA model prediction Label data.

3.3.2 select absolute error function as the evaluation index of typhoon tracks forecast model:

Wherein, N '_LFor R '₂₂Input data or label data total number, i '_LFor R '₂₂The serial number of middle data, 1≤i '_L ≤N′_L,For R '₂₂In the i-th '_LA true label data,It is the i-th '_LThe label data of a model prediction.

3.3.3 by R '₂Input mixing CNN-LSTM model, is trained mixing CNN-LSTM model, fitting obtains platform Wind path diameter prediction model, the loss function of model are root-mean-square error function, and evaluation index is absolute error function.Specific steps It is:

3.3.3.1 loop initialization number serial number n_c=1；

3.3.3.2 initializing the number of iterations epoch=10；Epoch value interval is 10~k_e, and be positive integer；k_eIt is all Positive integer, and k_e≥10；

3.3.3.3 initialization batch size batchsize=32；The value interval of batchsize is 32~N_F, and be 32 Positive integer times；

3.3.3.4 initializing learning rate lr=0.1；The value interval of lr isk_lIt is just whole between 1~9 Number.

3.3.3.5 model.compile (loss, optimizer (lr), the metrics) function built in Keras is called Compiling mixing CNN-LSTM model, the parameter setting of function are as follows: loss function loss is root-mean-square error function, optimization algorithm Optimizer is Adam algorithm, and learning rate lr is lr, and evaluation index metrics is that evaluation index is MAE.

3.3.3.6 by call Keras built in model.fit (data_train, labels_train, epochs, Batch_size) function mixes CNN-LSTM model to be fitted, and obtains the loss function value loss_value1 of training set.By letter Several parameter setting are as follows: training data data_train is R '₂₁N_LA input data, training label data labels_train For R '₂₁N_LA label data, the number of iterations epochs are epoch, and crowd size batch_size is batchsize.

3.3.3.7 by calling model.evaluate (data_test, the labels_test) function built in Keras Assessment mixing CNN-LSTM model, obtains the MAE of mixing CNN-LSTM model.By the parameter setting of function are as follows: test data Data_test is R '₂₂N_L' a input data, test data labels_test are R '₂₂N_L' a label data.

If 3.3.3.8 epoch≤k_eAnd batchsize≤N_FAndRecord n-th_cThe epoch of secondary circulation, Batchsize, lr, the loss function value loss_value2 of training set and the ACC of test set are called built in Keras Model.save (filename) function saves n_cSecondary to recycle obtained model, filename is set as model name.Enable n_c=n_c+ 1, lr=lr × 0.1, turn 3.3.3.5；If epoch≤k_eAnd batchsize≤N_FButEnable batchsize= Batchsize+32 turns 3.3.3.4；If epoch≤k_eBut batchsize > N_F, epoch=epoch+1 is enabled, 3.3.3.2 is turned； If epoch > k_e, turn 3.3.3.9.

3.3.3.9 by the 1st, 2.., n_c..., epoch, batchsize, lr, loss_ that C circulation is recorded Value2, MAE are integrated into the complete documentation of a model of fit.Searched out from complete documentation the smallest loss_value2 and Cycle-index serial number n ' corresponding to the smallest MAE_c.By the n-th '_cThe epoch of secondary preservation, batchsize, lr is as optimal ginseng Number, by the n-th '_cThe model of secondary preservation is as typhoon tracks forecast model.

3.4 utilize R '₃Training mixing CNN-LSTM model, fitting obtain intensity of typhoon prediction model:

3.4.1 select root-mean-square error function as the loss function of intensity of typhoon prediction model:

Wherein, loss function value is loss_value3, N_LFor R '₂₁Input data or label data total number, i_LFor R′₃₁The serial number of middle data, 1≤i_L≤N_L,For R '₃₁In i-th_LA true label data,It is i-th_LA model prediction Label data；

3.4.2 select absolute error function as the evaluation index of intensity of typhoon prediction model:

Wherein, N '_LFor R '₃₂Input data or label data total number, i '_LFor R '₃₂The serial number of middle data, 1≤i '_L ≤N′_L,For R '₂₂In the i-th '_LA true label data,It is the i-th '_LThe label data of a model prediction；

3.4.3 parameter regulation means are used, by calling the built-in function of Keras, by R '₃Input mixing CNN-LSTM mould Type is trained mixing CNN-LSTM model, and fitting obtains intensity of typhoon prediction model, and the loss function of model is root mean square Error function, evaluation index are absolute error function.It comprises the concrete steps that:

3.4.3.1 loop initialization number serial number n_c=1；

3.4.3.2 initializing the number of iterations epoch=10；Epoch value interval is 10~k_e, and be positive integer；k_eIt is all Positive integer, and k_e≥10；

3.4.3.3 initialization batch size batchsize=32；The value interval of batchsize is 32~N_F, and be 32 Positive integer times；

3.4.3.4 initializing learning rate lr=0.1；The value interval of lr isk_lIt is just whole between 1~9 Number.

3.4.3.5 model.compile (1oss, optimizer (lr), the metrics) function built in Keras is called Compiling mixing CNN-LSTM model, the parameter setting of function are as follows: loss function loss is root-mean-square error function, optimization algorithm Optimizer is Adam algorithm, and learning rate lr is lr, and evaluation index metrics is that evaluation index is MAE.

3.4.3.6 by call Keras built in model.fit (data_train, labels_train, epochs, Batch_size) function mixes CNN-LSTM model to be fitted, and obtains the loss function value loss_value3 of training set.By letter Several parameter setting are as follows: training data data_train is R '₃₁N_LA input data, training label data labels_train For R '₃₁N_LA label data, the number of iterations epochs are epoch, and crowd size batch_size is batchsize.

3.4.3.7 by calling model.evaluate (data_test, the labels_test) function built in Keras Assessment mixing CNN-LSTM model, obtains the MAE of assessment models.By the parameter setting of function are as follows: test data data_test is R′₃₂N_L' a input data, test data labels_test are R '₃₂N_L' a label data.

If 3.4.3.8 epoch≤k_eAnd batchsize≤N_FAndRecord n-th_cThe epoch of secondary circulation, Batchsize, lr, the loss function value loss_value3 of training set and the MAE of test set are called built in Keras Model.save (filename) function saves n_cSecondary to recycle obtained model, filename is set as model name.Enable n_c=n_c+ 1, lr=lr × 0.1, turn 3.4.3.5；If epoch≤k_eAnd batchsize≤N_FButEnable batchsize= Batchsize+32 turns 3.4.3.4；If epoch≤k_eBut batchsize > N_F, epoch=epoch+1 is enabled, 3.4.3.2 is turned； If epoch > k_e, turn 3.4.3.9.

3.4.3.9 by the 1st, 2 ..., n_c..., epoch, batchsize, lr, loss_ that C circulation is recorded Value3, MAE are integrated into the complete documentation of a model of fit.Searched out from complete documentation the smallest loss_value3 and Cycle-index serial number n ' corresponding to the smallest MAE_c.By the n-th '_cThe epoch that secondary circulation saves, batchsize, lr is as optimal Parameter, by the n-th '_cThe model that secondary circulation saves is as intensity of typhoon prediction model.

4th step is predicted that method is using three kinds of typhoon forecast models

4.1 enable the longitude and latitude range of target prediction region A for U_A°×W_A°, U_A° be A longitude span, W_A° be A latitude Span, it is desirable that the tropical cyclone record region of A and first step selection belongs to same sea area.The data type of selected downloading is A In existing highest resolution u_A°×u_A° gridding Atmosphere and Ocean variable data.Therefore, total lattice point number G=N × M in A, Wherein

4.2 consistent predicted times when selecting with training pattern walk k, and look-back time walks b.

Typhoon forecast is carried out after Atmosphere and Ocean variable data corresponding to lattice point in 4.3 downloading A, method is:

4.3.1 coordinate is established as origin using first point of the two-dimentional longitude and latitude grid of A, enabling origin is (1,1)；

4.3.2 initialization ordinate is variable n=1；

4.3.3 initialization abscissa is variable m=1；

4.3.4 use the method downloading coordinate of downloading lattice point Atmosphere and Ocean variable data for the big Qihai of lattice point at (m, n) Foreign variable data X_{[m, n]}.Method is:

According to the current time t of prediction₁And at (m, n) lattice point longitude station lon_{[m, n]}, Position Latitude lat_{[m, n]}It adopts The single-point variable data method for down loading described in step 1.3.3, downloads t from weather forecast center one by one₁, t₁- 6, t₁- 12, ...t₁The Atmosphere and Ocean variable data of lattice point, is recorded as respectively at -6b moment (m, n)Each lattice point is one in grid Five-tuple structure, five-tuple structure are (NUM, LOC, TT, X^P, X^S), wherein NUM is the coordinate (m, n) of grid, and LOC is (m, n) Locate the longitude and latitude value of lattice point, TT is the set of a time, contains t₁, t₁- 6, t₁- 12 ..., t₁The time at -6b moment, X^PIt is the atmospheric variable data at lattice point TT each moment at (m, n), X^SIt is the ocean variable data at lattice point TT each moment at (m, n), X^P With X^SIt is the Atmosphere and Ocean data X of lattice point at (m, n) after synthesis_{[m, n]}。

If 4.3.5 n≤N and m≤M, by X_{[m, n]}It is input to whether typhoon forms prediction model, lattice point 6k at prediction (m, n) Whether there is typhoon formation after hour.If prediction result is not formed, m=m+1 is enabled, 4.3.4 is turned；If prediction result is to be formed, turn 4.3.6；If n≤N but m > M, enable n=n+1, turn 4.3.3；If n > M, turn the 5th step；

4.3.6 initialization prediction sequence number variable n_f=1；

4.3.7 current according to the prediction of real-time update using the method for 4.3.4 step downloading lattice point Atmosphere and Ocean variable data Moment t₂And the longitude and latitude position of lattice point re-downloads data and obtains X ' at (m, n)_{[m, n]}, the X ' that will obtain_{[m, n]}Input typhoon Path prediction model, obtains n-th_fSecondary path prediction result As typhoon is in t₂+6k×n_fThe longitude and latitude position at moment.

4.3.8 by X '_{[m, n]}Intensity of typhoon prediction model is inputted, obtains n-th_fSecondary prediction of strength result As platform Wind is in t₂+6k×n_fThe intensity value at moment.

4.3.9 ifIt is not 0, the longitude and latitude position obtained according to predictionIt searches for corresponding with latitude and longitude value identical in A Lattice point coordinate (m ', n '), enable n_f=n_f+ 1, turn 4.3.7.IfIt is 0, n_f=prediction total degree NF, turns 4.3.10.

4.3.10 by n_fSecondary typhoon tracks forecastConnection is from t after drawing₂From moment 6k×n_fPredicted path after hour, n_fSecondary intensity of typhoon predictionAs corresponding path On intensity of typhoon value.

5th step terminates.

Following beneficial effect can be reached using the present invention:

The invention proposes a kind of typhoon forecast methods based on deep learning mixing CNN-LSTM model.This method uses The Atmosphere and Ocean variable data of history tropical cyclone optimal path data set and the whole world is constructed and has been fitted based on mixing convolution mind The typhoon forecast model of interacting depth learning model through network (CNN) and shot and long term Memory Neural Networks (LSTM), to typhoon Whether path and intensity after being formed and being formed are predicted.The model can automatically extract atmosphere in conjunction with 3DCNN and 2DCNN The space characteristics of extra large table variable data, and the time serial message during typhoon occurrence and development is automatically extracted in conjunction with LSTM. Eventually by the various temporal correlations conducive to typhoon occurrence and development variable of capture to reach higher predictablity rate.Using this Invention can obtain following effect:

1. the present invention can be convenient using existing public data collection and deep learning frame carry out model building and Prediction.The first step of the present invention obtains existing public data collection from tropical cyclone official website and weather forecast center and is used for model The preparation of building and training, second step carry out the building of model using deep learning frame Keras.Therefore cost is small, and easily In implementation.

2. the present invention numerical prediction prediction technique, statistical prediction methods and traditional machine learning different from the past Prediction technique.The present invention automatically extracts the ability of the non-linear relation between variable using deep learning and CNN and LSTM exists Ability in studying space data and time series data constructs the deep learning model for being directed to typhoon.The present invention is second Step and third step are trained the model of building using the typhoon data and Atmosphere and Ocean variable data of recent decades, therefore 4th step can directly effectively improve the accuracy rate of typhoon forecast using trained model.

3. different prediction targets can be arranged by different demands in the present invention, scalability is strong.Third step of the present invention is logical Training fitting typhoon formation model is crossed, typhoon track model and intensity of typhoon model, the 4th step are utilized respectively these three models Successfully whether typhoon can be formed, the position and intensity after formation are predicted.

Detailed description of the invention

Fig. 1 is overview flow chart of the present invention.

Fig. 2 is the flow chart of second step building mixing CNN-LSTM model of the present invention.

Fig. 3 is the building-block of logic of the mixing CNN-LSTM model of second step building of the present invention.

Specific embodiment

The first step obtains tropical cyclone optimal path data set, selects tropical cyclone optimal path record, matching record pair The Atmosphere and Ocean variable data answered, data set needed for constructing training pattern, method is:

1.1 from NOAA, (National Oceanic and Atmospheric Administration, is translated into: state, the U.S. Family ocean and Atmospheric Administration) NCDC (National Centers for environmental information, is translated Are as follows: U.S. environment information centre) websitehttps://www.ncdc.noaa.gov/ibtracs/It obtains 1979~2016 years and passes through The IBTrACs of WMO (World Meteorological Organization, be translated into World Meteorological Organization) certification (International Best Track Archive for Climate Stewardship is translated into international weather management most Good file store) version tropical cyclone optimal path data set.It selects to press tropical cyclone in tropical cyclone optimal path data set Generate the tropical cyclone optimal path data set (element is tropical cyclone optimal path record) in the region that sea area is divided.Often Tropical cyclone optimal path in one region tropical cyclone optimal path data set including different year records, and wraps again in each year Include the tropical cyclone optimal path record of different item numbers.Single tropical cyclone optimal path record containing the tropical cyclone from formed to All Time, longitude and latitude position and the strength information of extinction.

The 1.2 a certain fixed area for enabling needs detect a typhoon are Western Pacific (West Pacific, abbreviation WP), the heat of WP Band cyclone optimal path data set is Q, checks whether the maximum intensity in Q on single tropical cyclone optimal path reaches one by one 64kt (32.7m/s) puts the single tropical cyclone record (i.e. typhoon record) that maximum intensity reaches 64kt to typhoon data set Q₁, by maximum intensity reach 34kt (17.2m/s) be not up to 64kt tropical cyclone single record (i.e. tropical storm record) put To tropical storm data set Q₂, thus construct typhoon data set Q₁, tropical storm data set Q₂, remaining is unsatisfactory for the note of condition Record is cast out.

1.3 downloading Q₁Middle whole typhoon records the time of all moment record points, longitude and latitude position, is become using Atmosphere and Ocean Data set generation method is measured according to Q₁Middle whole typhoon records the time of all moment record points, center longitude position carries out group It knits, Q needed for obtaining model training₁Corresponding Atmosphere and Ocean variable data collectionMethod is:

1.3.1 by Q₁In typhoon record carry out it is randomly ordered, enable N₁For Q₁Typhoon record number, initialization typhoon note Record serial number n_p=1；

1.3.2 enablingIt is n-th_pThe record point number of typhoon record, initialization n-th_pThe record point sequence of typhoon record Number

1.3.3 according to n-th_pThe of article typhoon recordThe time of a record pointLongitude station Position LatitudeUsing single-point variable data method for down loading, from ECMWF (European Centre for Medium- Range Weather Forecasts, is translated into European Center for Medium Weather Forecasting) downloading ERA-Interim is analyzed again on website As corresponding Atmosphere and Ocean variable data, single-point variable data method for down loading is data:

1.3.3.1 the range of Atmosphere and Ocean variable data downloading is set as with n-th_pThe of article typhoon recordA record point Longitude and latitude position centered on 10 ° × 10 ° longitude and latitude grids, 10 ° expression data cover longitude span or latitude span.

1.3.3.2 the resolution ratio of Atmosphere and Ocean variable data downloading is set as n-th_pThe of article typhoon recordA record point 0.125 ° × 0.125 ° of the highest resolution that can be got.0.125 ° of longitude interval or latitude interval indicated between lattice point, area Domain A₁Interior lattice point number is 81 × 81.

1.3.3.3 the atmospheric variable of downloading is set as n-th_pThe of article typhoon recordZonal wind (the zonal of a record Wind, abbreviation u), meridional wind (meridional wind, abbreviation v), and geopotential unit (Geopotential Height, referred to as Z), temperature (Temperature, abbreviation T), relative humidity (Relative Humidity, abbreviation Rh)；Set the atmosphere of downloading Layer where the variable space is n-th_pThe of article typhoon recordRetrievable the 1000/975/925/850/800/ of a record The shell of compression of 700/600/500/400/300/200/100hpa.

1.3.3.4 the ocean variable of downloading is set as n-th_pThe of article typhoon recordSea surface temperature (the Sea of a record Surface Temperature, abbreviation SST)；Layer where setting the ocean variable space of downloading is n-th_pThe of article typhoon recordA record obtains one layer of extra large table.

1.3.3.5 n-th is downloaded_pThe of article typhoon recordThe corresponding atmospheric variable data X of a record point^PAnd ocean Variable data X^S, by X^PAnd X^SIt is packaged composition n-th_pThe of article typhoon recordThe time of a record point and longitude and latitude are big accordingly Qihai ocean variable data, is denoted as It is n-th_pThe of article typhoon recordA note The time of point is recorded,It is n-th_pThe of article typhoon recordThe longitude of a record point,It is n-th_pTyphoon The of recordThe latitude of a record point.

1.3.4 enabling

If 1.3.5 n_p≤N₁AndTurn 1.3.3；If n_p≤N₁ButEnable n_p=n_p+ 1, it goes to step 1.3.2；If n_p> N₁, execute 1.3.6.

1.3.6 will

Packing group becomes a data set, as Q₁Corresponding Atmosphere and Ocean variable data collectionContaining N₁Item platform The corresponding Atmosphere and Ocean variable data of all record points of wind.

1.4 downloading Q₂Whole tropical storms record the time of all moment record points, center longitude position, using step The 1.3 Atmosphere and Ocean variable data set creation methods, according to Q₂Whole tropical storms record all moment record point time, Longitude and latitude position carries out tissue, Q needed for obtaining model training₂Corresponding Atmosphere and Ocean variable data collection

1.5 according to Q₁, Q₂AndThe data set of different prediction models is trained for different prediction target buildings, Method is:

If 1.5.1 prediction target is whether typhoon forms, data needed for whether training typhoon forms prediction model are constructed Collect R₁, method is:

1.5.1.1 according to Q₁AndConstruct the required data set R of training pattern₁₁, method is:

1.5.1.1.1 typhoon record number n is initialized_p=1；Initialize R₁₁Input data serial number q=1；

1.5.1.1.2 from Q₁In select n-th_pTyphoon record；FromIn select n-th_pTyphoon records corresponding atmosphere Ocean variable dataIt is packaged, is recorded as

1.5.1.1.3 n-th is enabled_pThe numbered positions of record point of first intensity greater than 64kt are in typhoon record

1.5.1.1.4 selected predicted time T=24h, predicted time walk k=T/6=4；Selected look-back time walks b=2.

1.5.1.1.5 ifFromIn pick out n-th_pThe in article record The corresponding Atmosphere and Ocean variable data of a record point is packaged, composition n-th_pTyphoon recording needle is to training pattern institute structure The input data built, is recorded as X '_q, X '_qIndicate R₁₁In q-th of input data；Q=q+1 is enabled, 1.5.1.1.6 is turned；IfDirectly turn 1.5.1.1.6.

1.5.1.1.6 n is enabled_p=n_p+ 1, if n_p≤N₁, turn 1.5.1.1.2；If n_p> N₁, illustrate to have obtainedM₁For R₁₁In input data number, M₁≤N₁, turn 1.5.1.1.7.

1.5.1.1.7 willIt is packaged, is utilized Q₁In all typhoon recording needles to training pattern institute The input data set of building.By M₁A input data tagged 1, is utilized Q₁For label constructed by training pattern Data set.Input data set and label data collection have collectively constituted data set R needed for whether trained typhoon forms model₁₁, R₁₁ Containing M₁A input data and M₁A label data.

1.5.1.2 according to Q₂AndConstruct the required data set R of training pattern₁₂, method is:

1.5.1.2.1 N is enabled₂For Q₂Tropical storm record number, initialize tropical storm record number m_p=1；Initialization R₁₂Input data serial number h=1；

1.5.1.2.2 from Q₂In select m_pTropical storm record, enables m_pThe number of tropical storm record record point ForFromIn select m_pTropical storm records corresponding Atmosphere and Ocean variable dataIt is packaged, note For

1.5.1.2.3 enabling m_pThe numbered positions of the record point of first maximum of intensity are in tropical storm record

1.5.1.2.4 selected predicted time T=24h, predicted time walk k=T/6=4；Selected look-back time walks b=2.

1.5.1.2.5 ifFromIn pick out m_pThe in article record The corresponding Atmosphere and Ocean variable data of a record point is packaged, and forms m_pTropical storm recording needle is to training mould Input data constructed by type, is recorded as X '_h, X '_hIndicate R₁₂In h-th of input data；H=h+1 is enabled, 1.5.1.2.6 is turned；IfDirectly turn 1.5.1.2.6.

1.5.1.2.6 m is enabled_p=m_p+ 1, if m_p≤N₂, turn 1.5.1.2.2；If m_p> N₂, illustrate to have obtainedM₂≤N₂, M₂For R₁₂In input data number, turn 1.5.1.2.7.

1.5.1.2.7 willIt is packaged, is utilized Q₂In all tropical storm recording needles to training mould Input data set constructed by type.By M₂A input data tagged 0, is utilized Q₂For constructed by training pattern Label data collection.Input data set and label data collection have collectively constituted data set needed for whether trained typhoon forms model R₁₂, R₁₂Containing M₂A input data and M₂A label data.

1.5.1.3 by R₁₁And R₁₂Merge into data set R needed for whether trained typhoon forms prediction model₁, R₁Containing model Required I₁A input data, I₁A label data, I₁=M₁+M₂；Turn 1.6.

If 1.5.2 prediction target is typhoon track, data set R needed for constructing training typhoon track prediction model₂, method It is: to Q₁In all typhoons record latitude and longitude value and corresponding Atmosphere and Ocean variable dataDislocation processing is carried out, at dislocation The method of reason is:

1.5.2.1 typhoon record number n is initialized_p=1；Enable I₂Indicate R₂In input data number, initialization data set R₂The serial number i of middle input data_p=1；

1.5.2.2 from Q₁In at random select n-th_pTyphoon record.

1.5.2.3 n-th is enabled_pThe numbered positions recorded a little in typhoon record areInitializationThis is enabled to record Above the total number of all record points is

1.5.2.4 predicted time T=24h is selected, predicted time walks k=T/6=4；Selected look-back time walks b=2.

1.5.2.5 ifFromA record point starts, selection Totally 3 The Atmosphere and Ocean variable data of record point, which is packaged, is used as R₂In i-th_pA input data, is recorded asSelection theRecord The typhoon longitude and latitude position of point is as R₂In i-th_pThe label data of a input data, is recorded asTurn 1.5.2.6；IfButSelection The Atmosphere and Ocean variable data of totally 3 record points As R₂In i-th_pThe input data of a sample, is recorded asIt willThe typhoon longitude and latitude of a record point Positional value is assigned to 0, and as R₂In i-th_pThe label data of a input data, is recorded asTurn 1.5.2.6；IfEnable n_p=n_p+ 1, turn 1.5.2.7.

1.5.2.6 enablingi_p=i_p+ 1, turn 1.5.2.5.

If 1.5.2.7 n_p≤N₁, turn 1.5.2.2；If n_p> N₁, turn 1.5.2.8.

1.5.2.8 willIt is packaged, is utilized Q₁In all typhoon recording needles to training pattern institute structure The path input data set built.It willFourth packet, is utilized Q₁In all typhoon recording needles to training pattern institute The path label data set of building.Path input data set and path label data set have collectively constituted trained typhoon tracks forecast Data set R needed for model₂, R₂Containing I₂A path input dataAnd I₂A path label data Path label data are latitude and longitude value；Turn 1.6.

If 1.5.3 prediction target is intensity of typhoon, to Q₁In all typhoons record intensity value and corresponding Atmosphere and Ocean Variable dataDislocation processing is carried out, data set R needed for constructing training intensity of typhoon prediction model₃, R₃In containing utilize Q₁ In all typhoon recording needles to intensity input data set constructed by training pattern and utilize Q₁In all typhoon recording needles to instruction Practice intensity label data set constructed by model, intensity input data concentration contains I₂A input data, I₂A input data isContain I in intensity label data set₂A label data, I₂A mark Signing data is Intensity label data is intensity value；Turn 1.6；

1.6 couples of R₁, R₂, R₃It is pre-processed；Respectively to R₁Input data set, R₂Input data set and R₂Number of tags According to collection, R₃Input data set and R₃Label data collection be normalized by formula (1),

Wherein R indicates R₁, R₂, R₃In any one data set input data set or R₂, R₃In any one data set Label data collection, R_maxIndicate the maximum value of the data set, R_minIndicate the minimum value of the data set.R₁, R₂, R₃It is pretreated Data set is respectively R '₁, R '₂, R '₃；

Second step, building deep learning mix CNN-LSTM model, and construction method is as shown in Figure 2:

2.1 build deep learning environment.Deep learning library Keras based on Python is installed on the server, and is used Google deep learning frame TensorFlow is as rear end.

The 2.2 deep learning library Keras building deep learnings based on Python mix CNN-LSTM typhoon forecast model, side Method is:

2.2.1 learn the 3DCNN module of atmospheric variable space characteristics in building mixing CNN-LSTM model.In Fig. 3 Shown in 3DCNN module:

The input layer for being 1 layer of Keras by the configuration settings of neural network in 3DCNN module, 3 (m_A=3) layer Three dimensional convolution Layer, 3 (m_A=3) layer batch regularization layer, 1 layer of tiling layer, 1 layer of full articulamentum.It sets first layer Three dimensional convolution layer and contains 323 dimensions convolution kernel (i.e. filter), the size of convolution kernel isStep-length is Setting Second layer Three dimensional convolution layer contains3 a dimensions convolution kernel (i.e. filter), the size of convolution kernel areStep-length is Setting Third layer Three dimensional convolution layer contains3 a dimensions convolution kernel (i.e. filter), the size of convolution kernel areStep-length isEvery layer three Tie up one layer of batch regularization layer of addition after convolutional layer, and the activation primitive that all layers of neuron selects is ' Relu ' function, specifically Form is Relu (x)=max (0, x) (taking the maximum value in 0 and x).Set the neuron number d of full articulamentum_A=100 It is a.

2.2.2 learn the 2DCNN module of ocean variable space feature in building mixing CNN-LSTM model.In Fig. 3 Shown in 2DCNN module:

The input layer for being 1 layer of Keras by the configuration settings of neural network in 3DCNN module, 3 (m_B=3) layer two-dimensional convolution Layer, 3 (m_B=3) layer batch regularization, 1 layer of tiling layer and 1 layer of full articulamentum composition.The 1st layer of two-dimensional convolution layer is set to containA 2D convolution kernel, the size of convolution kernel are Step-length isThe 2nd layer of two-dimensional convolution layer is set to containA 2D convolution kernel, convolution kernel Having a size ofStep-length isThe 3rd layer of two-dimensional convolution layer is set to containA 2D convolution kernel, the size of convolution kernel areStep-length is After every layer of two-dimensional convolution layer plus one layer of batch regularization layer, and all layers of neuron selection Activation primitive is ' Relu ' function.The neuron number of full articulamentum is set as d_B=100.

2.2.3 learn the LSTM module of temporal aspect in building mixing CNN-LSTM model.Such as LSTM module institute in Fig. 3 Show:

The structure of neural network in LSTM module is set as the LSTM of 1 layer of Keras layer.The neuron number that LSTM is arranged is d_C=100, the activation primitive that all layers of neuron selects is still ' Relu ' function.

2.2.4 by 3DCNN module, 2DCNN module, LSTM module is by with the Sequential progressive die built in Keras Block connection, and one layer of output layer is added, output layer neuron number is set as 1, is mapped as single value to export, neuron The activation primitive selected is ' sigmoid ' function σ (x), and concrete form is(e is the bottom of natural logrithm Number), the model after connection is to mix CNN-LSTM model.Wherein, the input of 3DCNN module is R '₁, R '₂, R '₃Middle input number According to the atmospheric variable data of collection, export as airspace feature vector 0^P, airspace feature vector O^PVector length is full connection The neuron number d of layer_A=100.The input of 2DCNN module is R '₁, R '₂, R '₃The ocean variable data of middle input data set, Output is ocean space feature vector O^S, ocean space feature vector O^SVector length is the neuron number d of full articulamentum_B= 100.The input of LSTM module is merging vector, that is, Atmosphere and Ocean space characteristics of 3DCNN module and 2DCNN module output vector Vector O=[O^P, O^S], merging vector O is got by (Merge layers) merging of merging layer of Keras, and the length of O is d_A+d_B= 200.The output of LSTM module is 1 output mapping value.

Third step, be fitted to obtain by constructed data set training mixing CNN-LSTM model typhoon whether formed it is pre- Model, typhoon tracks forecast model, intensity of typhoon prediction model are surveyed, method is:

3.1 according to demand by R '₁, R '₂, R '₃In data upset at random after (input data and label data need one a pair of Answer), the ratio cut partition by 7: 3 is training set R '₁₁, R '₂₁, R '₃₁With test set R '₁₂, R '₂₂, R '₃₂。R′₁₁With R '₁₂Contain N respectively_F And N_F' a input data and label data, R '₂₁With R '₂₂Contain N respectively_LAnd N_L' a input data and label data, R '₃₁With R '₃₂ Equally contain N respectively_LAnd N_L' a input data and label data.

3.2 utilize R '₁Training mixing CNN-LSTM model obtains whether typhoon forms prediction model to fitting, and method is such as Under:

3.2.1 it selects to intersect loss function of the entropy function as mixing CNN-LSTM model, functional form is as follows:

Wherein, loss function value is loss_valuel, N_FFor R '₁₁Input data or label data total number, i_FFor R′₁₁The serial number of middle data, 1≤i_F≤N_F,For R '₁₁In i-th_FA true label data,For the number of tags of model prediction According to,Expression pairTake logarithm.

3.2.2 evaluation index of the accuracy rate (ACC) as model quality is selected:

Wherein, N_AFor R '₁₂The middle sum by the correct label data of model prediction, N_F' it is R '₁₂Label data sum.

3.2.3 parameter regulation means are used, by calling the built-in function of Keras, by R '₁Input mixing CNN-LSTM mould Type is trained mixing CNN-LSTM model, and fitting obtains the no formation prediction model of typhoon, and the loss function of model is to intersect Entropy function, evaluation index ACC.It comprises the concrete steps that:

3.2.3.1 loop initialization number serial number nc=1；

3.2.3.2 initializing the number of iterations epoch=10；Epoch value interval is 10~100, and is positive integer；

3.2.3.3 initialization batch size batchsize=32；The positive integer times that the value of batchsize is 32, and need small In N_F；

3.2.3.4 initializing learning rate lr=0.1；The value interval of lr is 0.1~10^-6；

3.2.3.5 model.compile (loss, optimizer (lr), the metrics) function built in Keras is called Compiling mixing CNN-LSTM model, the parameter setting of function are as follows: loss function loss is to intersect entropy function, optimization algorithm Optimizer is Adam algorithm, and learning rate lr is lr, and evaluation index metrics is ACC.

3.2.3.6 by call Keras built in model.fit (data_train, labels_train, epochs, Batch size) function come be fitted mixing CNN-LSTM model, obtain the loss function value loss_value1 of training set.By letter Several parameter setting are as follows: training data data_train is R '₁₁N_FA input data, training label data labels_train For R '₁₁N_FA label data, the number of iterations epochs are epoch, and crowd size batch_size is batchsize.

3.2.3.7 by calling model.evaluate (data_test, the labels_test) function built in Keras Assessment mixing CNN-LSTM model, obtains the ACC of mixing CNN-LSTM model.By the parameter setting of function are as follows: test data Data_test is R '₁₂N_F' a input data, test data labels_test are R '₁₂N_F' a label data.

If 3.2.3.8 epoch≤100 and batchsize≤N_FAnd lr >=10^-6, record n-th_cThe epoch of secondary circulation, Batchsize, lr, the loss function value loss_value1 of training set and the ACC of test set are called built in Keras Model.save (filename) function saves n_cSecondary to recycle obtained model, filename is set as model name.Enable n_c=n_c+ 1, lr=lr × 0.1, turn 3.2.3.5；If epoch≤100 and batchsize≤N_FBut lr < 10^-6, enable batchsize= Batchsize+32 turns 3.2.3.4；If epoch≤100 but batchsize > N_F, epoch=epoch+1 is enabled, is turned 3.2.3.2；If epoch > 100, turns 3.2.3.9.

3.2.3.9 by the 1st, 2 ..., n_c..., epoch, batchsize, lr, loss_ that C circulation is recorded Value1, ACC are integrated into the complete documentation of a model of fit.Searched out from complete documentation the smallest loss_value1 and Cycle-index serial number n ' corresponding to maximum ACC_c.By the n-th '_cThe epoch that secondary circulation saves, batchsize, lr is as optimal Parameter, by the n-th '_cWhether the model that secondary circulation saves forms prediction model as typhoon.

3.3 utilize R '₂Training mixing CNN-LSTM model, fitting obtain typhoon tracks forecast model, the method is as follows:

3.3.1 select root-mean-square error function as the loss function of typhoon tracks forecast model:

Wherein, loss function value is loss_value2, N_LFor R '₂₁Input data or label data total number, i_LFor R′₂₁The serial number of middle data, 1≤i_L≤N_L,For R '₂₁In i-th_LA true label data,It is i-th_LA model prediction Label data.

3.3.2 select absolute error function as the evaluation index of typhoon tracks forecast model:

Wherein, N '_LFor R '₂₂Input data or label data total number, i '_LFor R '₂₂The serial number of middle data, 1≤i '_L ≤N′_L,For R '₂₂In the i-th '_LA true label data,It is the i-th '_LThe label data of a model prediction.

3.3.3 by R '₂Input mixing CNN-LSTM model, is trained mixing CNN-LSTM model, fitting obtains platform Wind path diameter prediction model, the loss function of model are root-mean-square error function, and evaluation index is absolute error function.Specific steps It is:

3.3.3.1 loop initialization number serial number n_c=1；

3.3.3.2 initializing the number of iterations epoch=10；Epoch value interval is 10~100, and is positive integer.

3.3.3.3 initialization batch size batchsize=32；The positive integer times that the value of batchsize is 32, and need small In N_L。

3.3.3.4 initializing learning rate lr=0.1；The value interval of lr is 0.1~10^-6。

3.3.3.5 model.compile (1oss, optimizer (lr), the metrics) function built in Keras is called Compiling mixing CNN-LSTM model, the parameter setting of function are as follows: loss function loss is root-mean-square error function, optimization algorithm Optimizer is Adam algorithm, and learning rate lr is lr, and evaluation index metrics is that evaluation index is MAE.

3.3.3.6 by call Keras built in model.fit (data_train, labels_train, epochs, Batch_size) function mixes CNN-LSTM model to be fitted, and obtains the loss function value loss_value2 of training set.By letter Several parameter setting are as follows: training data data_train is R '₂₁N_LA input data, training label data labels_train For R '₂₁N_LA label data, the number of iterations epochs are epoch, and crowd size batch_size is batchsize.

3.3.3.7 by calling model.evaluate (data_test, the labels_test) function built in Keras Assessment mixing CNN-LSTM model, obtains the MAE of mixing CNN-LSTM model.By the parameter setting of function are as follows: test data Data_test is R '₂₂N_L' a input data, test data labels_test are R '₂₂N_L' a label data can be obtained The MAE of assessment models.

If 3.3.3.8 epoch≤100 and batchsize≤N_FAnd lr >=10^-6, the epoch of n-th c times circulation is recorded, Batchsize, lr, the loss function value loss_value2 of training set and the ACC of test set are called built in Keras Model.save (filename) function saves n_cSecondary to recycle obtained model, filename is set as model name.Enable n_c=n_c+ 1, lr=lr × 0.1, turn 3.3.3.5；If epoch≤100 and batchsize≤N_FBut lr < 10^-6, enable batchsize= Batchsize+32 turns 3.3.3.4；If epoch≤100 but batchsize > N_F, epoch=epoch+1 is enabled, is turned 3.3.3.2；If epoch > 100, turns 3.3.3.9.

3.3.3.9 by the 1st, 2 ..., n_c..., epoch, batchsize, lr, loss_ that C circulation is recorded Value2, MAE are integrated into the complete documentation of a model of fit.Searched out from complete documentation the smallest loss_value2 and Cycle-index serial number n ' corresponding to the smallest MAE_c.By the n-th '_cThe epoch that secondary circulation saves, batchsize, lr is as optimal Parameter, by the n-th '_cThe model that secondary circulation saves is as typhoon tracks forecast model.

3.4 utilize R '₃Training mixing CNN-LSTM model, fitting obtain intensity of typhoon prediction model:

3.4.1 select root-mean-square error function as the loss function of intensity of typhoon prediction model:

Wherein, loss function value is loss_value3, N_LFor R '₂₁Input data or label data total number, i_LFor R′₃₁The serial number of middle data, 1≤i_L≤N_L,For R '₃₁In i-th_LA true label data,It is i-th_LA model prediction Label data；

3.4.2 select absolute error function as the evaluation index of intensity of typhoon prediction model:

Wherein, N '_LFor R '₃₂Input data or label data total number, i '_LFor R '₃₂The serial number of middle data, 1≤i '_L ≤N′_L,For R '₂₂In the i-th '_LA true label data,It is the i-th '_LThe label data of a model prediction；

3.4.3 parameter regulation means are used, by calling the built-in function of Keras, by R '₃Input mixing CNN-LSTM mould Type is trained mixing CNN-LSTM model, and fitting obtains intensity of typhoon prediction model, and the loss function of model is root mean square Error function, evaluation index are absolute error function.It comprises the concrete steps that:

3.4.3.1 loop initialization number serial number n_c=1；

3.4.3.2 initializing the number of iterations epoch=10；Epoch value interval is 10~100, and is positive integer.

3.4.3.3 initialization batch size batchsize=32；The positive integer times that the value of batchsize is 32, and need small In N_L。

3.4.3.4 initializing learning rate lr=0.1；The value interval of lr is 0.1~10^-6。

3.4.3.5 model.compile (loss, optimizer (lr), the metrics) function built in Keras is called Compiling mixing CNN-LSTM model, the parameter setting of function are as follows: loss function loss is root-mean-square error function, optimization algorithm Optimizer is Adam algorithm, and learning rate lr is lr, and evaluation index metrics is that evaluation index is MAE.

3.4.3.6 by call Keras built in model.fit (data_train, labels_train, epochs, Batch_size) function mixes CNN-LSTM model to be fitted, and obtains the loss function value loss_value3 of training set.By letter Several parameter setting are as follows: training data data_train is R '₃₁N_LA input data, training label data labels_train For R '₃₁N_LA label data, the number of iterations epochs are epoch, and crowd size batch_size is batchsize.Thus may be used Obtain the loss function value loss_value3 of training set.

3.4.3.7 by calling model.evaluate (data_test, the labels_test) function built in Keras Assessment mixing CNN-LSTM model, obtain the MAE of mixing CNN-LSTM model.By the parameter setting of function are as follows: test data Data_test is R '₃₂N_L' a input data, test data labels_test are R '₃₂N_L' a label data can be obtained The MAE of assessment models.

3.4.3.8 epoch≤100 and batchsize≤N_FAnd lr >=10^-6, record n-th_cThe epoch of secondary circulation, Batchsize, lr, the loss function value loss_value3 of training set and the MAE of test set are called built in Keras Model.save (filename) function saves n_cSecondary to recycle obtained model, filename is set as model name.Enable n_c=n_c+ 1, lr=lr × 0.1, turn 3.4.3.5；If epoch≤100 and batchsize≤N_FBut lr < 10^-6, enable batchsize= Batchsize+32 turns 3.4.3.4；If epoch≤100 but batchsize > N_F, epoch=epoch+1 is enabled, is turned 3.4.3.2；If epoch > 100, turns 3.4.3.9.

3.4.3.9 by the 1st, 2 ..., n_c..., epoch, batchsize, lr, loss_ that C circulation is recorded Value3, MAE are integrated into the complete documentation of a model of fit.Searched out from complete documentation the smallest loss_value3 and Cycle-index serial number n ' corresponding to the smallest MAE_c.By the n-th '_cThe epoch that secondary circulation saves, batchsize, lr is as optimal Parameter, by the n-th '_cThe model that secondary circulation saves is as intensity of typhoon prediction model.

4th step carries out typhoon forecast, method using three kinds of typhoon forecast models fixed sea area A a certain to Western Pacific It is:

4.1 enable the longitude and latitude range of A for U_A°×W_A°, U_A° be A longitude span, W_A° be A latitude span.Under selected The data type of load is the gridding Atmosphere and Ocean variable data of 0.125 ° × 0.125 ° of the existing highest resolution of A.Therefore, A Interior total lattice point number G=N × M, wherein

4.2 consistent predicted times when selecting with training pattern walk k=4, and look-back time walks b=2.

Typhoon forecast is carried out after Atmosphere and Ocean variable data corresponding to lattice point in 4.3 downloading A, method is:

4.3.1 coordinate is established as origin using first point of the two-dimentional longitude and latitude grid of A, enabling origin is (1,1)；

4.3.2 initialization ordinate is variable n=1；

4.3.3 initialization abscissa is variable m=1；

4.3.4 use the method downloading coordinate of downloading lattice point Atmosphere and Ocean variable data for the big Qihai of lattice point at (m, n) Foreign variable data X_{[m, n]}.Method is:

According to the current time t of prediction₁And at (m, n) lattice point longitude station lon_{[mm, n]}, Position Latitude lat_{[m, n]}It adopts The single-point variable data method for down loading described in step 1.3.3, downloads t from weather forecast center one by one₁, t₁- 6, t₁- 12 moment The Atmosphere and Ocean variable data of lattice point, is recorded as respectively at (m, n)Each lattice point is one five in grid Tuple structure, five-tuple structure are (NUM, LOC, TT, X^P, X^S), wherein NUM is the coordinate (m, n) of grid, and LOC is at (m, n) The longitude and latitude value of lattice point, TT are the set of a time, contain t₁, t₁- 6, t₁The time at -12 moment, X^PIt is at (m, n) The atmospheric variable data at lattice point TT each moment, X^SIt is the ocean variable data at lattice point TT each moment at (m, n), X^PWith X^sAfter synthesis The Atmosphere and Ocean data X of lattice point as at (m, n)_{[m, n]}。

If 4.3.5 n≤N and m≤M, by X_{[m, n]}It is input to whether typhoon forms prediction model, lattice point 24 at prediction (m, n) Whether there is typhoon formation after hour.If prediction result is not formed, m=m+1 is enabled, 4.3.4 is turned；If prediction result is to be formed, turn 4.3.6；If n≤N but m > M, enable n=n+1, turn 4.3.3；If n > M, turn the 5th step；

4.3.6 initialization prediction sequence number variable n_f=1；

4.3.7 current according to the prediction of real-time update using the method for 4.3.4 step downloading lattice point Atmosphere and Ocean variable data Moment t₂And the longitude and latitude position of lattice point re-downloads data and obtains X ' at (m, n)_{[m, n]}, the X ' that will obtain_{[m, n]}Input typhoon Path prediction model, obtains n-th_fSecondary path prediction result As typhoon is in t₂+24×n_fThe longitude and latitude position at moment.

4.3.8 by X '_{[m, n]}Intensity of typhoon prediction model is inputted, obtains n-th_fSecondary prediction of strength result As platform Wind is in t₂+24×n_fThe intensity value at moment.

4.3.9 ifIt is not 0, the longitude and latitude position obtained according to predictionIt searches for corresponding with latitude and longitude value identical in A Lattice point coordinate (m ', n '), enable n_f=n_f+ 1, turn 4.3.7.IfIt is 0, n_f=prediction total degree NF, turns 4.3.10.

4.3.10 by n_fSecondary typhoon tracks forecastConnection is from t after drawing₂Moment Play 24 × n_fPredicted path after hour, n_fSecondary intensity of typhoon predictionAs corresponding road Intensity of typhoon value on diameter.

5th step terminates.

Claims (10)

1. a kind of typhoon forecast method based on deep learning mixing CNN-LSTM model, it is characterised in that the following steps are included:

The first step obtains tropical cyclone optimal path data set, selects tropical cyclone optimal path record, what matching record pair was answered Atmosphere and Ocean variable data, data set needed for constructing training pattern, method is:

1.1 download history tropical cyclone optimal path data set from tropical cyclone official website, select tropical cyclone optimal path The tropical cyclone optimal path data set in the region that sea area is divided is generated in data set by tropical cyclone, tropical cyclone is best The element that path data integrates is tropical cyclone optimal path record；Include in each region tropical cyclone optimal path data set The tropical cyclone optimal path of different year records, and each year includes the tropical cyclone optimal path record of different item numbers again, single Tropical cyclone optimal path record is believed containing the tropical cyclone from being formed the All Time of extinction, longitude and latitude position and intensity Breath；

The 1.2 a certain fixed area for enabling needs detect a typhoon are region A₁, A₁Tropical cyclone optimal path data set be Q, one by one It checks whether the maximum intensity in Q on single tropical cyclone optimal path reaches 64kt i.e. 32.7m/s, maximum intensity is reached The single tropical cyclone record of 64kt is that typhoon record is put to typhoon data set Q₁, maximum intensity is reached into 34kt and is not up to 64kt Tropical cyclone single record be tropical storm record put to tropical storm data set Q₂, thus construct typhoon data set Q₁, heat Band storm data set Q₂, remaining record for being unsatisfactory for condition casts out；

1.3 downloading Q₁Middle whole typhoon records the time of all moment record points, longitude and latitude position, using Atmosphere and Ocean variable number According to set creation method according to Q₁Middle whole typhoon records the time of all moment record points, center longitude position carries out tissue, Q needed for obtaining model training₁Corresponding Atmosphere and Ocean variable data collectionMethod is:

1.3.1 by Q₁In typhoon record carry out it is randomly ordered, enable N₁For Q₁Typhoon record number, initialization typhoon record sequence Number n_p=1；

1.3.2 enablingIt is n-th_pThe record point number of typhoon record, initialization n-th_pThe record point serial number of typhoon record

1.3.3 according to n-th_pThe of article typhoon recordThe time of a record pointLongitude stationLatitude PositionN-th is downloaded from weather forecast center website using single-point variable data method for down loading_pThe of article typhoon recordThe corresponding atmospheric variable data X of a record point^PAnd ocean variable data X^S, by X^PAnd X^SIt is packaged composition n-th_pTyphoon The of recordThe time of a record point and the corresponding Atmosphere and Ocean variable data of longitude and latitude, are denoted asIt is n-th_pThe of article typhoon recordThe time of a record point, It is n-th_pThe of article typhoon recordThe longitude of a record point,It is n-th_pThe of article typhoon recordA record point Latitude；

1.3.4 enabling

If 1.3.5 n_p≤N₁AndTurn 1.3.3；If n_p≤N₁ButEnable n_p=n_p+ 1, it goes to step 1.3.2；If n_p> N₁, execute 1.3.6；

1.3.6 will Packing group As a data set, as Q₁Corresponding Atmosphere and Ocean variable data collectionContaining N₁All record points pair of typhoon The Atmosphere and Ocean variable data answered；

1.4 downloading Q₂Whole tropical storms record the time of all moment record points, center longitude position, using step 1.3 step The Atmosphere and Ocean variable data set creation method, according to Q₂Whole tropical storms record the time of all moment record points, warp Position Latitude carries out tissue, Q needed for obtaining model training₂Corresponding Atmosphere and Ocean variable data collection

1.5 according to Q₁, Q₂AndFor the data set of the different different prediction models of prediction target building training, method It is:

If 1.5.1 prediction target is whether typhoon forms, data set R needed for whether training typhoon forms prediction model is constructed₁, Method is:

1.5.1.1 according to Q₁AndConstruct the required data set R of training pattern₁₁, R₁₁In containing utilize Q₁In all typhoons Recording needle is to input data set constructed by training pattern and utilizes Q₁In all typhoon recording needles pair, M₁≤N₁, 1≤q≤M₁, By M₁A input data tagged 1, as label data collection；

1.5.1.2 according to Q₂AndConstruct the required data set R of training pattern₁₂, R₁₂In containing utilize Q₂In all torrid zones Storm recording needle is to input data set constructed by training pattern and utilizes Q₂In all tropical storms, M₂≤N₂, 1≤h≤M₂, N₂ For Q₂Tropical storm record number, by M₂A input data tagged 1, as label data collection；

1.5.1.3 by R₁₁And R₁₂Merge into data set R needed for whether trained typhoon forms prediction model₁, R₁Containing needed for model I₁A input data, I₁A label data, I₁=M₁+M₂；Turn 1.6；

If 1.5.2 prediction target is typhoon track, to Q₁In all typhoons record latitude and longitude value and corresponding Atmosphere and Ocean variable DataDislocation processing is carried out, data set R needed for constructing training typhoon track prediction model₂, R₂In containing utilize Q₁Middle institute There is typhoon recording needle to path input data set constructed by training pattern and utilizes Q₁In all typhoon recording needles to training mould Path label data set constructed by type, path input data concentration contain I₂A input data, I₂A input data is1≤i_p≤I₂, contain I in path label data set₂A label data, I₂A label data isPath label data are latitude and longitude value；Turn 1.6；

If 1.5.3 prediction target is intensity of typhoon, to Q₁In all typhoons record intensity value and corresponding Atmosphere and Ocean variable number According toDislocation processing is carried out, data set R needed for constructing training intensity of typhoon prediction model₃, R₃In containing utilize Q₁In own Typhoon recording needle is to intensity input data set constructed by training pattern and utilizes Q₁In all typhoon recording needles to training pattern Constructed intensity label data set, intensity input data concentration contain I₂A intensity input data, I₂A intensity input data is1≤i_p≤I₂, contain I in intensity label data set₂A label data, I₂A label data is Intensity label data is intensity value；Turn 1.6；

1.6 couples of R₁, R₂, R₃It is pre-processed: respectively to R₁Input data set, R₂Input data set and R₂Label data Collection, R₃Input data set and R₃Label data collection be normalized by formula (1),

Wherein R indicates R₁, R₂, R₃In any one data set input data set or R₂, R₃In any one data set label Data set, R_maxIndicate the maximum value of the data set, R_minIndicate the minimum value of the data set；R₁, R₂, R₃Pretreated data Collection is respectively R '₁, R '₂, R '₃；

Second step, building deep learning mix CNN-LSTM model, and method is:

2.1 build deep learning environment: installing the deep learning library Keras based on Python on the server, and use Google deep learning frame TensorFlow is as rear end；

The 2.2 deep learning library Keras building deep learnings based on Python mix CNN-LSTM typhoon forecast model, method It is:

2.2.1 learn the 3DCNN module of atmospheric variable space characteristics in building mixing CNN-LSTM model: will be in 3DCNN module The configuration settings of neural network are the input layer of 1 layer of Keras, m_ALayer Three dimensional convolution layer, m_ALayer batch regularization layer, 1 layer of tiling layer, 1 The full articulamentum of layer；The 1st layer of Three dimensional convolution layer is set to contain3 a dimension convolution kernels, the size of convolution kernel are Step-length isThe 2nd layer of Three dimensional convolution layer is set to contain3 a dimension convolution kernels, the size of convolution kernel areStep-length isSet m_ALayer Three dimensional convolution layer contains3 a dimension convolution kernels, The size of convolution kernel isStep-length isOne is added after every layer of Three dimensional convolution layer Layer batch regularization layer；The neuron number of full articulamentum is set as d_AIt is a, wherein parameter It is positive integer；

2.2.2 learn the 2DCNN module of ocean variable space feature in building mixing CNN-LSTM model: will be in 2DCNN module The configuration settings of neural network are the input layer of 1 layer of Keras, m_BLayer two-dimensional convolution layer, m_BLayer batch regularization, 1 layer of tiling layer and 1 The full articulamentum composition of layer；The 1st layer of two-dimensional convolution layer is set to containA 2D convolution kernel, the size of convolution kernel areStep-length ForThe 2nd layer of two-dimensional convolution layer is set to containA 2D convolution kernel, the size of convolution kernel areStep-length isEvery layer of two dimension volume ..., setting m_BLayer two-dimensional convolution layer containsA 2D convolution kernel, the size of convolution kernel areStep-length isAfter every layer of two-dimensional convolution layer plus one layer is criticized regularization layer；Set the mind of full articulamentum It is d through first number_BIt is a, whereinIt is positive integer；

2.2.3 learn the LSTM module of temporal aspect: neural network in setting LSTM module in building mixing CNN-LSTM model Structure be 1 layer of Keras LSTM layer, the neuron number of LSTM is set for d_CIt is a；

2.2.4 by 3DCNN module, 2DCNN module, LSTM module with the Sequential sequent modular built in Keras by being connected It connects, and adds one layer of output layer, set output layer neuron number is 1, model after connection as mixing CNN-LSTM model； Wherein, the input of 3DCNN module is R '₁, R '₂, R '₃The atmospheric variable data of middle input data set export as airspace feature Vector O^P, airspace feature vector O^PVector length is the neuron number d of full articulamentum_A；The input of 2DCNN module is R ′₁, R '₂, R '₃The ocean variable data of middle input data set exports as airspace feature vector O^S, airspace feature vector O^SVector length is the neuron number d of full articulamentum_B；The input of LSTM module be 3DCNN module and 2DCNN module export to The merging vector O=[O of amount^P, O^S], merge vector O be it is laminated and get by the merging of Keras, the length of O is d_A+d_B；LSTM The output of module is 1 output mapping value；

Third step is fitted to obtain whether typhoon forms prediction mould by constructed data set training mixing CNN-LSTM model Type, typhoon tracks forecast model, intensity of typhoon prediction model, method is:

3.1 according to demand under the premise of input data and label data correspond by R '₁, R '₂, R '₃In data beat at random It disorderly, is training set R ' by a: b ratio cut partition₁₁, R '₂₁, R '₃₁With test set R '₁₂, R '₂₂, R '₃₂, wherein a, b are positive integer and a ≥b；R′₁₁With R '₁₂Contain N respectively_FAnd N_F' a input data and label data, R '₂₁With R '₂₂Contain N respectively_LAnd N_L' a input data And label data, R '₃₁With R '₃₂Equally contain N respectively_LAnd N_L' a input data and label data；

3.2 utilize R '₁Training mixing CNN-LSTM model obtains whether typhoon forms prediction model to fitting, the method is as follows:

3.2.1 it selects to intersect loss function of the entropy function as mixing CNN-LSTM model, functional form is as follows:

Wherein, loss function value is loss_value1, N_FFor R '₁₁Input data or label data total number, i_FFor R '₁₁ The serial number of middle data, 1≤i_F≤N_F,For R '₁₁In i-th_FA true label data,For the label data of model prediction,Expression pairTake logarithm；

3.2.2 evaluation index of the accuracy rate (ACC) as model quality is selected:

Wherein, N_AFor R '₁₂The middle sum by the correct label data of model prediction, N_F' it is R '₁₂Label data sum；

3.2.3 by R '₁Input mixing CNN-LSTM model, is trained mixing CNN-LSTM model, fitting obtains the no shape of typhoon At prediction model, the loss function of model is to intersect entropy function, evaluation index ACC；

3.3 utilize R '₂Training mixing CNN-LSTM model, fitting obtain typhoon tracks forecast model, the method is as follows:

3.3.1 select root-mean-square error function as the loss function of typhoon tracks forecast model:

Wherein, loss function value is loss_value2, N_LFor R '₂₁Input data or label data total number, i_LFor R '₂₁ The serial number of middle data, 1≤i_L≤N_L,For R '₂₁In i-th_LA true label data,It is i-th_LThe label of a model prediction Data；

3.3.2 select absolute error function as the evaluation index of typhoon tracks forecast model:

Wherein, N '_LFor R '₂₂Input data or label data total number, i '_LFor R '₂₂The serial number of middle data, 1≤i '_L≤N ′_L,For R '₂₂In the i-th '_LA true label data,It is the i-th '_LThe label data of a model prediction；

3.3.3 by R '₂Input mixing CNN-LSTM model, is trained mixing CNN-LSTM model, fitting obtains typhoon track Prediction model, the loss function of model are root-mean-square error function, and evaluation index is absolute error function；

3.4 utilize R '₃Training mixing CNN-LSTM model, fitting obtain intensity of typhoon prediction model:

3.4.1 select root-mean-square error function as the loss function of intensity of typhoon prediction model:

Wherein, loss function value is loss_value3, N_LFor R '₂₁Input data or label data total number, i_LFor R '₃₁ The serial number of middle data, 1≤i_L≤N_L,For R '₃₁In i-th_LA true label data,It is i-th_LThe label of a model prediction Data；

3.4.2 select absolute error function as the evaluation index of intensity of typhoon prediction model:

Wherein, N '_LFor R '₃₂Input data or label data total number, i '_LFor R '₃₂The serial number of middle data, 1≤i '_L≤N ′_L,For R '₂₂In the i-th '_LA true label data,It is the i-th '_LThe label data of a model prediction；

3.4.3 by R '₃Input mixing CNN-LSTM model, is trained mixing CNN-LSTM model, fitting obtains intensity of typhoon Prediction model, the loss function of model are root-mean-square error function, and evaluation index is absolute error function；

4th step is predicted that method is using three kinds of typhoon forecast models

4.1 enable the longitude and latitude range of target prediction region A for U_A°×W_A°, U_A° be A longitude span, W_A° be A latitude span, It is required that the tropical cyclone record region of A and first step selection belongs to same sea area；The data type of selected downloading is to have in A Highest resolution u_A°×u_A° gridding Atmosphere and Ocean variable data.Therefore, total lattice point number G=N × M in A, wherein

4.2 consistent predicted times when selecting with training pattern walk k, and look-back time walks b；

Typhoon forecast is carried out after Atmosphere and Ocean variable data corresponding to lattice point in 4.3 downloading A, method is:

4.3.1 coordinate is established as origin using first point of the two-dimentional longitude and latitude grid of A, enabling origin is (1,1)；

4.3.2 initialization ordinate is variable n=1；

4.3.3 initialization abscissa is variable m=1；

4.3.4 the method downloading coordinate of downloading lattice point Atmosphere and Ocean variable data is used to become for the Atmosphere and Ocean of lattice point at (m, n) Measure data X_{[m, n]}；Method is: according to the current time t of prediction₁And at (m, n) lattice point longitude station lon_{[m, n]}, latitude position Set lat_{[m, n]}Using single-point variable data method for down loading described in step 1.3.3, t is downloaded one by one from weather forecast center₁, t₁- 6, t₁- 12 ... t₁The Atmosphere and Ocean variable data of lattice point, is recorded as respectively at -6b moment (m, n)Each lattice point is one in grid Five-tuple structure, five-tuple structure are (NUM, LOC, TT, X^P, X^S), wherein NUM is the coordinate (m, n) of grid, and LOC is (m, n) Locate the longitude and latitude value of lattice point, TT is the set of a time, contains t₁, t₁- 6, t₁- 12 ..., t₁The time at -6b moment, X^PIt is the atmospheric variable data at lattice point TT each moment at (m, n), X^SIt is the ocean variable data at lattice point TT each moment at (m, n), X^P With X^SIt is the Atmosphere and Ocean data X of lattice point at (m, n) after synthesis_{[m, n]}；

If 4.3.5 n≤N and m≤M, by X_{[m, n]}It is input to whether typhoon forms prediction model, predicts at (m, n) lattice point 6k hours Whether there is typhoon formation afterwards；If prediction result is not formed, m=m+1 is enabled, 4.3.4 is turned；If prediction result is to be formed, turn 4.3.6；If n≤N but m > M, enable n=n+1, turn 4.3.3；If n > M, turn the 5th step；

4.3.6 initialization prediction sequence number variable n_f=1；

4.3.7 using the method for 4.3.4 step downloading lattice point Atmosphere and Ocean variable data, according to the prediction current time of real-time update t₂And the longitude and latitude position of lattice point re-downloads data and obtains X ' at (m, n)_{[m, n]}, the X ' that will obtain_{[m, n]}Input typhoon track Prediction model obtains n-th_fSecondary path prediction resultAs typhoon is in t₂+6k×n_fThe longitude and latitude position at moment；

4.3.8 by X '_{[m, n]}Intensity of typhoon prediction model is inputted, obtains n-th_fSecondary prediction of strength resultAs typhoon exists t₂+6k×n_fThe intensity value at moment；

4.3.9 ifIt is not 0, the longitude and latitude position obtained according to predictionSearch for lattice corresponding with latitude and longitude value identical in A Point coordinate (m ', n '), enables n_f=n_f+ 1, turn 4.3.7；IfIt is 0, n_f=prediction total degree NF, turns 4.3.10；

4.3.10 by n_fSecondary typhoon tracks forecastConnection is from t after drawing₂6k from moment × n_fPredicted path after hour, n_fSecondary intensity of typhoon predictionOn as corresponding path Intensity of typhoon value；

5th step terminates.

2. the typhoon forecast method as described in claim 1 based on deep learning mixing CNN-LSTM model, it is characterised in that 1.3.3 it walks described according to n-th_pThe of article typhoon recordThe time of a record pointLongitude stationLatitude PositionUsing single-point variable data method for down loading, from weather forecast center website, corresponding big Qihai is downloaded in selection The specific method of foreign variable data is:

1.3.3.1 the range of Atmosphere and Ocean variable data downloading is set as with n-th_pThe of article typhoon recordThe warp of a record point W ° × w ° longitude and latitude grid centered on Position Latitude, w are the positive integer less than 180, the longitude span of w ° of expression data cover Or latitude span；

1.3.3.2 the resolution ratio of Atmosphere and Ocean variable data downloading is set as n-th_pThe of article typhoon recordA record point can obtain U ° × u ° of the highest resolution got；The u ° of longitude interval or latitude interval indicated between lattice point, region A₁Interior lattice point number is

1.3.3.3 the atmospheric variable of downloading is set as n-th_pThe of article typhoon recordThe variable of a record, is denoted as p₁, p₂..., p_i..., p_I, I is atmospheric variable species number, and I is positive integer, and selected variable is that atmosphere relevant to typhoon occurrence and development becomes Amount；Layer is n-th where setting the atmospheric variable space of downloading_pThe of article typhoon recordThe retrievable l of a record₁, l₂..., l_i..., l_zLayer, Z are the atmospheric variable number of plies, and Z is positive integer；

1.3.3.4 the ocean variable of downloading is set as n-th_pThe of article typhoon recordThe variable of a record, is denoted as s₁, s₂..., s_i..., s_J, J is ocean variable species number, and J is positive integer, and selected variable is that ocean relevant to typhoon occurrence and development becomes Amount；Layer where setting the ocean variable space of downloading is n-th_pThe of article typhoon recordThe retrievable f of a record₁, f₂..., f_i... f_WLayer, W are the ocean variable number of plies, and W is positive integer；

1.3.3.5 n-th is downloaded_pThe of article typhoon recordThe corresponding atmospheric variable data X of a record point^PAnd ocean variable Data X^S, by X^PAnd X^SIt is packaged composition n-th_pThe of article typhoon recordIt is a to record the time put and longitude and latitude big Qihai accordingly Foreign variable data, is denoted as

3. the typhoon forecast method as described in claim 1 based on deep learning mixing CNN-LSTM model, it is characterised in that 1.5.1.1 it walks described according to Q₁AndConstruct the required data set R of training pattern₁₁Method be:

1.5.1.1.1 typhoon record number n is initialized_p=1；Initialize R₁₁Input data serial number q=1；

1.5.1.1.2 from Q₁In select n-th_pTyphoon record；FromIn select n-th_pTyphoon records corresponding Atmosphere and Ocean Variable data

It is packaged, is recorded as

1.5.1.1.3 n-th is enabled_pThe numbered positions of record point of first intensity greater than 64kt are in typhoon record

1.5.1.1.4 it selectes predicted time and walks k, look-back time walks b, and enabling k=T/6, T is predicted time, and is 6 positive integer Times, any positive integer of the b between 1~k；

1.5.1.1.5 ifFromIn pick out n-th_pThe in article record The corresponding Atmosphere and Ocean variable data of a record point is packaged, composition n-th_pTyphoon record For input data constructed by training pattern, it is recorded as X '_q, X '_qIndicate R₁₁In q-th of input data, enable q=q+1, turn 1.5.1.1.6；IfDirectly turn 1.5.1.1.6；

1.5.1.1.6 n is enabled_p=n_p+ 1, if n_p≤N₁, turn 1.5.1.1.2；If n_p> N₁, illustrate to have obtained M₁For R₁₁In input data number, M₁≤N₁, turn 1.5.1.1.7；

1.5.1.1.7 willIt is packaged, is utilized Q₁In all typhoon recording needles to constructed by training pattern Input data set；By M₁A input data tagged 1, is utilized Q₁For label data collection constructed by training pattern, Input data set and label data collection have collectively constituted data set R needed for whether trained typhoon forms model₁₁, R₁₁Containing M₁It is a defeated Enter data and M₁A label data.

4. the typhoon forecast method as described in claim 1 based on deep learning mixing CNN-LSTM model, it is characterised in that 1.5.1.2 it walks described according to Q₂AndConstruct the required data set R of training pattern₁₂Method be:

1.5.1.2.1 tropical storm record number m is initialized_p=1；Initialize R₁₂Input data serial number h=1:

1.5.1.2.2 from Q₂In select m_pTropical storm record, enables m_pThe number of tropical storm record record point isFromIn select m_pTropical storm records corresponding Atmosphere and Ocean variable dataIt is packaged, note For

1.5.1.2.3 enabling m_pThe numbered positions of the record point of first maximum of intensity are in tropical storm record

1.5.1.2.4 it selectes predicted time and walks k, look-back time walks b, and enabling k is T/6, and T is predicted time, and is 6 positive integer Times, any positive integer of the b between 1~k；

1.5.1.2.5 ifFromIn pick out m_pThe in article record The corresponding Atmosphere and Ocean variable data of a record point is packaged, and forms m_pTropical storm Recording needle is recorded as X ' to input data constructed by training pattern_h, X '_hIndicate R₁₂In h-th of input data, enable h=h+1, Turn 1.5.1.2.6；IfDirectly turn 1.5.1.2.6；

1.5.1.2.6 m is enabled_p=m_p+ 1, if m_p≤N₂, turn 1.5.1.2.2；If m_p> N₂, illustrate to have obtained M₂For R₁₂In input data number, M₂≤N₂, turn 1.5.1.2.7；

1.5.1.2.7 willIt is packaged, is utilized Q₂In all tropical storm recording needles to training pattern institute structure The input data set built；By M₂A input data tagged 0, is utilized Q₂For number of tags constructed by training pattern According to collection；Input data set and label data collection have collectively constituted data set R needed for whether trained typhoon forms model₁₂, R₁₂Contain M₂A input data and M₂A label data.

5. the typhoon forecast method as described in claim 1 based on deep learning mixing CNN-LSTM model, it is characterised in that 1.5.2 it walks described to Q₁In all typhoons record latitude and longitude value and corresponding Atmosphere and Ocean variable data X_Q1Carry out dislocation processing Method be:

1.5.2.1 typhoon record number n is initialized_p=1；Enable I₂Indicate R₂In input data number, initialization data set R₂In The serial number i of input data_p=1；

1.5.2.2 from Q₁In at random select n-th_pTyphoon record；

1.5.2.3 n-th is enabled_pThe numbered positions recorded a little in typhoon record areInitializationIt enables this record to own Recording the total number put is

1.5.2.4 it selectes predicted time and walks k, look-back time walks b, k T/6, the positive integer times that T is 6, b appointing between 1~k One positive integer；

1.5.2.5 ifFromA record point starts, selection Total b+1 The Atmosphere and Ocean variable data of a record point, which is packaged, is used as R₂In i-th_pA input data, is recorded asSelection theThe typhoon longitude and latitude position of point is recorded as R₂In i-th_pThe label data of a input data, is recorded asTurn 1.5.2.6；IfButSelectionTotal b+1 record The Atmosphere and Ocean variable data of point is as R₂In i-th_pThe input data of a sample, is recorded asIt willThe typhoon longitude and latitude position value of a record point is assigned to 0, and as R₂In i-th_pA input number According to label data, be recorded asTurn 1.5.2.6；IfEnable n_p=n_p+ 1, turn 1.5.2.7；

1.5.2.6 enablingi_p=i_p+ 1, turn 1.5.2.5；

If 1.5.2.7 n_p≤N₁, turn 1.5.2.2；If n_p> N₁, turn 1.5.2.8；

1.5.2.8 willIt is packaged, is utilized Q₁In all typhoon recording needles to constructed by training pattern Path input data set；It willIt is packaged, is utilized Q₁In all typhoon recording needles to constructed by training pattern Path label data set；Path input data set and path label data set have collectively constituted trained typhoon track prediction model Required data set R₂, R₂Containing I₂A path input data and I₂A path label data.

6. the typhoon forecast method as described in claim 1 based on deep learning mixing CNN-LSTM model, it is characterised in that Described in second step when building deep learning mixing CNN-LSTM model, 3DCNN module, 2DCNN module, institute in LSTM module are constructed The activation primitive for having the neuron of layer to select is ' Relu ' function, and ' Relu ' function concrete form is Relu (x)=max (0, x), Take the maximum value in 0 and x.

7. the typhoon forecast method as described in claim 1 based on deep learning mixing CNN-LSTM model, it is characterised in that 2.2.4 walking the activation primitive that the neuron of the output layer selects is ' sigmoid ' function σ (x), and concrete form isE is the truth of a matter of natural logrithm.

8. the typhoon forecast method as described in claim 1 based on deep learning mixing CNN-LSTM model, it is characterised in that 3.2.3 it walks the use parameter regulation means to be trained mixing CNN-LSTM model, fitting obtains the no formation prediction of typhoon Model comprises the concrete steps that:

3.2.3.1 loop initialization number serial number n_c=1；

3.2.3.2 initializing the number of iterations epoch=10；Epoch value interval is 10~k_e, and be positive integer；k_eIt is whole with being positive Number, and k_e≥10；

3.2.3.3 initialization batch size batchsize=32；The value of batchsize is 32 positive integer times, and need to be less than N_F；

3.2.3.4 initializing learning rate lr=0.1；The value interval of lr isk_lFor the positive integer between 1~9；

3.2.3.5 model.compile (loss, optimizer (lr), the metrics) function built in Keras is called to compile Mix CNN-LSTM model, the parameter setting of function are as follows: loss function loss is to intersect entropy function, optimization algorithm optimizer For Adam algorithm, learning rate lr is lr, and evaluation index metrics is ACC；

3.2.3.6 by calling model.fit (data_train, labels_train, epochs, batch_ built in Keras Size) function mixes CNN-LSTM model to be fitted, and obtains the loss function value loss_value1 of training set；By the ginseng of function Number setting are as follows: training data data_train is R '₁₁N_FA input data, training label data labels_train are R '₁₁ N_FA label data, the number of iterations epochs are epoch, and crowd size batch_size is batchsize；

3.2.3.7 it is assessed by calling model.evaluate (data_test, the labels_test) function built in Keras CNN-LSTM model is mixed, the ACC of mixing CNN-LSTM model is obtained；The input of function is set are as follows: test data data_ Test is R '₁₂N_F' a input data, test data labels_test are R '₁₂N_F' a label data；

If 3.2.3.8 epoch≤k_eAnd batchsize≤N_FAndRecord n-th_cThe epoch of secondary circulation, Batchsize, lr, the loss function value loss_value1 of training set and the ACC of test set are called built in Keras Model.save (filename) function saves n_cSecondary to recycle obtained model, filename is set as model name；Enable n_c=n_c+ 1, lr=lr × 0.1, turn 3.2.3.5；If epoch≤k_eAnd batchsize≤N_FButEnable batchsize= Batchsize+32 turns 3.2.3.4；If epoch≤k_eBut batchsize > N_F, epoch=epoch+1 is enabled, 3.2.3.2 is turned； If epoch > k_e, turn 3.2.3.9；

3.2.3.9 by the 1st, 2 ..., n_c..., epoch, batchsize, lr, loss_value1 that C circulation is recorded, ACC is integrated into the complete documentation of a model of fit；The smallest loss_value1 and maximum is searched out from complete documentation Cycle-index serial number n ' corresponding to ACC_c；By the n-th '_cThe epoch of secondary preservation, batchsize, lr is as optimized parameter, by n′_cWhether the model of secondary preservation forms prediction model as typhoon.

9. the typhoon forecast method as described in claim 1 based on deep learning mixing CNN-LSTM model, it is characterised in that 3.3.3 it walks described by R '₂Input mixing CNN-LSTM model, is trained mixing CNN-LSTM model, fitting obtains typhoon Path prediction model comprises the concrete steps that:

3.3.3.1 loop initialization number serial number n_c=1；

3.3.3.2 initializing the number of iterations epoch=10；Epoch value interval is 10~k_e, and be positive integer；k_eIt is whole with being positive Number, and k_e≥10；

3.3.3.3 initialization batch size batchsize=32；The value of batchsize is 32 positive integer times, and need to be less than N_L；

3.3.3.4 initializing learning rate lr=0.1；The value interval of lr isk_lFor the positive integer between 1~9；

3.3.3.5 model.compile (loss, optimizer (lr), the metrics) function built in Keras is called to compile Mix CNN-LSTM model, the parameter setting of function are as follows: loss function loss is root-mean-square error function, optimization algorithm Optimizer is Adam algorithm, and learning rate lr is lr, and evaluation index metrics is that evaluation index is MAE；

3.3.3.6 model.fit (data_train, labels_train, epochs, the batch_ built in Keras are called Size) function mixes CNN-LSTM model to be fitted, and obtains the loss function value loss_value2 of training set；By the ginseng of function Number setting are as follows: training data data_train is R '₂₁N_LA input data, training label data labels_train are R '₂₁ N_LA label data, the number of iterations epochs are epoch, and crowd size batch_size is batchsize；

3.3.3.7 model.evaluate (data_test, the labels_test) function built in Keras is called to assess mixing CNN-LSTM model obtains the MAE of mixing CNN-LSTM model；By the parameter setting of function are as follows: test data data_test is R′₂₂N_L' a input data, test data labels_test are R '₂₂N_L' a label data；

If 3.3.3.8 epoch≤k_eAnd batchsize≤N_FAndRecord n-th_cThe epoch of secondary circulation, Batchsize, lr, the loss function value loss_value2 of training set and the ACC of test set are called built in Keras Model.save (filename) function saves n_cSecondary to recycle obtained model, filename is set as model name；Enable n_c=n_c+ 1, lr=lr × 0.1, turn 3.3.3.5；If epoch≤k_eAnd batchsize≤N_FButEnable batchsize= Batchsize+32 turns 3.3.3.4；If epoch≤k_eBut batchsize > N_F, epoch=epoch+1 is enabled, 3.3.3.2 is turned； If epoch > k_e, turn 3.3.3.9；

3.3.3.9 by the 1st, 2 ..., n_c..., epoch, batchsize, lr, loss_value2 that C circulation is recorded, MAE is integrated into the complete documentation of a model of fit；The smallest loss_value2 and the smallest is searched out from complete documentation Cycle-index serial number n ' corresponding to MAE_c；By the n-th '_cThe epoch of secondary preservation, batchsize, lr is as optimized parameter, by n′_cThe model of secondary preservation is as typhoon tracks forecast model.

10. the typhoon forecast method as described in claim 1 based on deep learning mixing CNN-LSTM model, it is characterised in that 3.4.3 walking described by R '₃Input mixing CNN-LSTM model, is trained mixing CNN-LSTM model, fitting obtains typhoon Model To Describe Strength of Blended comprises the concrete steps that:

3.4.3.1 loop initialization number serial number n_c=1；

3.4.3.2 initializing the number of iterations epoch=10；Epoch value interval is 10~k_e, and be positive integer；k_eIt is whole with being positive Number, and k_e≥10；

3.4.3.3 initialization batch size batchsize=32；The value of batchsize is 32 positive integer times, and need to be less than N_L；

3.4.3.4 initializing learning rate lr=0.1；The value interval of lr isk_lFor the positive integer between 1~9；

3.4.3.5 model.compile (loss, optimizer (lr), the metrics) function built in Keras is called to compile Mix CNN-LSTM model, the parameter setting of function are as follows: loss function loss is root-mean-square error function, optimization algorithm Optimizer is Adam algorithm, and learning rate lr is lr, and evaluation index metrics is that evaluation index is MAE；

3.4.3.6 model.fit (data_train, labels_train, epochs, the batch_ built in Keras are called Size) function mixes CNN-LSTM model to be fitted, and obtains the loss function value loss_value3 of training set；By the ginseng of function Number setting are as follows: training data data_train is R '₃₁N_LA input data, training label data labels_train are R '₃₁ N_LA label data, the number of iterations epochs are epoch, and crowd size batch_size is batchsize；

3.4.3.7 model.evaluate (data_test, the labels_test) function built in Keras is called to assess mixing CNN-LSTM model obtains the MAE of mixing CNN-LSTM model.The parameter setting of function are as follows: test data data_test is R′₃₂N_L' a input data, test data labels_test are R '₃₂N_L' a label data:

If 3.4.3.8 epoch≤k_eAnd batchsize≤N_FAndRecord n-th_cThe epoch of secondary circulation, Batchsize, lr, the loss function value loss_value3 of training set and the MAE of test set are called built in Keras Model.save (filename) function saves n_cSecondary to recycle obtained model, filename is set as model name；Enable n_c=n_c+ 1, lr=lr × 0.1, turn 3.4.3.5；If epoch≤k_eAnd batchsize≤N_FButEnable batchsize= Batchsize+32 turns 3.4.3.4；If epoch≤k_eBut batchsize > N_F, epoch=epoch+1 is enabled, 3.4.3.2 is turned； If epoch > k_e, turn 3.4.3.9；

3.4.3.9 by the 1st, 2 ..., n_c..., epoch, batchsize, lr, loss_value3 that C circulation is recorded, MAE is integrated into the complete documentation of a model of fit；The smallest loss_value3 and the smallest is searched out from complete documentation Cycle-index serial number n ' corresponding to MAE_c；By the n-th '_cThe epoch that secondary circulation saves, batchsize, lr as optimized parameter, By the n-th '_cThe model that secondary circulation saves is as intensity of typhoon prediction model.