scholarly journals Monthly rainfall forecasting by a hybrid neural network of discrete wavelet transform and deep learning

2021 ◽  
Author(s):  
Xueyi You ◽  
Ming Wei
Keyword(s):  
Neural Network ◽  
Time Series ◽  
Deep Learning ◽  
Wavelet Transform ◽  
Discrete Wavelet Transform ◽  
Monthly Rainfall ◽  
Discrete Wavelet ◽  
Rainfall Time Series ◽  
Efficiency Coefficient ◽  
Rainfall Forecasting

Actual rainfall forecast is critical to the management and allocation of water resources. In recent years, deep learning has been proved to be superior to traditional forecasting methods when predicting rainfall time series with high temporal and spatial variability. In this study, the discrete wavelet transform (DWT) and two typical deep learning approaches, namely long-short term memory (LSTM) and dilated causal convolutional neural network (DCCNN), are integrated innovatively and the hybrid model (DWT-CLSTM-DCCNN) is used for monthly rainfall forecasting for the first time. Monthly rainfall time series of four major cities in China (Beijing, Tianjin, Chongqing and Guangzhou) are used as the dataset of DWT-CLSTM-DCCNN. Firstly, two methods of sample construction are used to train DWT-CLSTM-DCCNN and their effects on the model performance are analyzed. Then, LSTM and the dilated causal convolutional network (DCCNN) are established as the benchmark models, and their forecast accuracy is compared with that of DWT-CLSTM-DCCNN. From the results of the evaluation criteria such as mean absolute error (MAE), root mean squared error (RMSE) and Nash-Sutcliffe model efficiency coefficient (NSE) as well as the fitting curve for forecasted rainfall, it can be concluded that the DWT-CLSTM-DCCNN developed in this study outperforms the benchmark models in model accuracy, peak and mutational rainfall capturing ability. Compared with the previous studies, DWT-CLSTM-DCCNN is proven to be better peak capture and more suitable for long-term rainfall forecasting.

scholarly journals Long-term trend analysis of rainfall using hybrid Discrete Wavelet Transform (DWT) based Mann-Kendall tests in central Gujarat region, India

MAUSAM ◽  
2021 ◽  
Vol 71 (2) ◽  
pp. 209-224
Author(s):  
RAJANI NIRAV V ◽  
TIWARI MUKESH K ◽  
CHINCHORKAR S S
Keyword(s):  
Time Series ◽  
Wavelet Transform ◽  
Discrete Wavelet Transform ◽  
Trend Analysis ◽  
Time Scales ◽  
Series Data ◽  
Trend Test ◽  
Discrete Wavelet ◽  
Rainfall Time Series

Trend analysis has become one of the most important issues in hydro-meteorological variables study due to climate change and the focus given to it in the recent past from the scientific community. In this study, long-term trends of rainfall are analyzed in eight stations located in semi-arid central Gujarat region, India by considering time series data of 116 years (1901-2016). Discrete wavelet transform (DWT) as a dyadic arrangement of continuous wavelet transformation combined with the widely applied and acknowledged Mann-Kendall (MK) trend analysis method were applied for analysis of trend and dominant periodicities in rainfall time series at monthly, annual and monsoonal time scales. Initially, rainfall time series applied in this study were decomposed using DWT to generate sub-time series at high and low frequencies, before applying the MK trend test. Further, the Sequential Mann-Kendall (SQMK) test was also applied to find out the trend changing points. The result showed that at the monthly annual and monsoon time scales, the trends in rainfall were significantly decreasing in most of the station. The 4-month and 8-month components were found as dominant at the monthly time series and the 2-year and 4-year component were found as dominant at the monsoon time series, whereas the 2-year components were observed as dominant in the annual time scale.

Maximal Overlap Discrete Wavelet Transform, Graph Theory And Backpropagation Neural Network In Stock Market Forecasting

2018 ◽  
Vol 5 (1) ◽  
pp. 41-46
Author(s):  
Rosalina Rosalina ◽  
Hendra Jayanto
Keyword(s):  
Neural Network ◽  
Graph Theory ◽  
Wavelet Transform ◽  
Stock Market ◽  
Discrete Wavelet Transform ◽  
Stock Price ◽  
Moving Average ◽  
Discrete Wavelet ◽  
Stock Market Forecasting ◽  
Artificial Neural

The aim of this paper is to get high accuracy of stock market forecasting in order to produce signals that will affect the decision making in the trading itself. Several experiments by using different methodologies have been performed to answer the stock market forecasting issues. A traditional linear model, like autoregressive integrated moving average (ARIMA) has been used, but the result is not satisfactory because it is not suitable for model financial series. Yet experts are likely observed another approach by using artificial neural networks. Artificial neural network (ANN) are found to be more effective in realizing the input-output mapping and could estimate any continuous function which given an arbitrarily desired accuracy. In details, in this paper will use maximal overlap discrete wavelet transform (MODWT) and graph theory to distinguish and determine between low and high frequencies, which in this case acted as fundamental and technical prediction of stock market trading. After processed dataset is formed, then we will advance to the next level of the training process to generate the final result that is the buy or sell signals given from information whether the stock price will go up or down.

scholarly journals PERAMALAN DATA PENUMPANG KERETA API DENGAN MENGGUNAKAN MAXIMAL OVERLAP DISCRETE WAVELET TRANSFORM- RECURRENT NEURAL NETWORK (MODWT-RNN)

2019 ◽  
Vol 12 (2) ◽  
pp. 164
Author(s):  
Mira Andriyani ◽  
Subanar Subanar
Keyword(s):  
Neural Network ◽  
Wavelet Transform ◽  
Discrete Wavelet Transform ◽  
Recurrent Neural Network ◽  
Public Transportation ◽  
Accurate Result ◽  
Nonlinear Models ◽  
Forecasting Model ◽  
Discrete Wavelet ◽  
The Public

The train is one of the public transportation that is very popular because it is affordable and free of congestion. There is often a buildup of train passengers at the station so that it sometimes causes an accumulation of passengers at the station and makes the situation at the station to be not conducive. In order to avoid a buildup of passengers, forecasting the number of passengers can be done. Forecasting is determined based on data in previous times. Data of train passengers in Java (excluding Jabodetabek) forms a non-stationary and contains nonlinear relationships between the lags. One of the nonlinear models that can be used is Recurrent Neural Network (RNN). Before RNN modeling, Maximal Overlap Wavelet Transform (MODWT) was used to make data more stationary. Forecasting model of train passengers in Java excluding Jabodetabek, Indonesia using MODWT-RNN results forecasting with RMSE is 252.85, while RMSE of SARIMA and RNN are 434.97 and 320.48. These results indicate that the MODWT-RNN model gives a more accurate result than SARIMA and RNN.

Sign in / Sign up

Export Citation Format

Share Document