Grid search of multilayer perceptron based on the walk-forward validation methodology

Multilayer perceptron neural network is one of the widely used method for load forecasting. There are hyperparameters which can be used to determine the network structure and used to train the multilayer perceptron neural network model. This paper aims to propose a framework for grid search model based on the walk-forward validation methodology. The training process will specify the optimal models which satisfy requirement for minimum of accuracy scores of root mean square error, mean absolute percentage error and mean absolute error. The testing process will evaluate the optimal models along with the other ones. The results indicated that the optimal models have accuracy scores near the minimum values. The US airline passenger and Ho Chi Minh city load demand data were used to verify the accuracy and reliability of the grid search framework.

Download Full-text

Grid search of exponential smoothing method: a case study of Ho Chi Minh City load demand

Indonesian Journal of Electrical Engineering and Computer Science ◽

10.11591/ijeecs.v19.i3.pp1121-1130 ◽

2020 ◽

Vol 19 (3) ◽

pp. 1121

Author(s):

Ngoc Thanh Tran ◽

Le Van Dai

Keyword(s):

Information Criterion ◽

Absolute Error ◽

Exponential Smoothing ◽

Smoothing Method ◽

Ho Chi Minh City ◽

Percentage Error ◽

Grid Search ◽

Load Demand ◽

Exponential Smoothing Method

The exponential smoothing method is one of the widely used methods for load forecasting. The taxonomy of exponential smoothing method shows that its trend and seasonal component affect the results of exponential smoothing method. This paper proposed a framework for grid search with the optimal model of exponential smoothing method based on math formulas. The training process will specify the optimal models which satisfy requirement of minimum of akaike information criterion, accuracy scores of the root mean square error, mean absolute percentage error, and mean absolute error. The testing process will evaluate the accuracy scores between the optimal models and all other ones. The results indicated that the optimal models have accuracy scores near the minimum values. The load demand data collected in Ho Chi Minh City were used to verify the accuracy and reliability of the grid search framework.

Download Full-text

Detection and Severity Evaluation of Combined Rail Defects Using Deep Learning

Vibration ◽

10.3390/vibration4020022 ◽

2021 ◽

Vol 4 (2) ◽

pp. 341-356

Author(s):

Jessada Sresakoolchai ◽

Sakdirat Kaewunruen

Keyword(s):

Neural Network ◽

Machine Learning ◽

Deep Learning ◽

Mean Absolute Error ◽

Absolute Error ◽

Machine Learning Techniques ◽

Rolling Stock ◽

Raw Data ◽

Learning Techniques ◽

Combined Defects

Various techniques have been developed to detect railway defects. One of the popular techniques is machine learning. This unprecedented study applies deep learning, which is a branch of machine learning techniques, to detect and evaluate the severity of rail combined defects. The combined defects in the study are settlement and dipped joint. Features used to detect and evaluate the severity of combined defects are axle box accelerations simulated using a verified rolling stock dynamic behavior simulation called D-Track. A total of 1650 simulations are run to generate numerical data. Deep learning techniques used in the study are deep neural network (DNN), convolutional neural network (CNN), and recurrent neural network (RNN). Simulated data are used in two ways: simplified data and raw data. Simplified data are used to develop the DNN model, while raw data are used to develop the CNN and RNN model. For simplified data, features are extracted from raw data, which are the weight of rolling stock, the speed of rolling stock, and three peak and bottom accelerations from two wheels of rolling stock. In total, there are 14 features used as simplified data for developing the DNN model. For raw data, time-domain accelerations are used directly to develop the CNN and RNN models without processing and data extraction. Hyperparameter tuning is performed to ensure that the performance of each model is optimized. Grid search is used for performing hyperparameter tuning. To detect the combined defects, the study proposes two approaches. The first approach uses one model to detect settlement and dipped joint, and the second approach uses two models to detect settlement and dipped joint separately. The results show that the CNN models of both approaches provide the same accuracy of 99%, so one model is good enough to detect settlement and dipped joint. To evaluate the severity of the combined defects, the study applies classification and regression concepts. Classification is used to evaluate the severity by categorizing defects into light, medium, and severe classes, and regression is used to estimate the size of defects. From the study, the CNN model is suitable for evaluating dipped joint severity with an accuracy of 84% and mean absolute error (MAE) of 1.25 mm, and the RNN model is suitable for evaluating settlement severity with an accuracy of 99% and mean absolute error (MAE) of 1.58 mm.

Download Full-text

Novel Spatiotemporal Feature Extraction Parallel Deep Neural Network for Forecasting Confirmed Cases of Coronavirus Disease 2019

10.1101/2020.04.30.20086538 ◽

2020 ◽

Author(s):

Chiou-Jye Huang ◽

Yamin Shen ◽

Ping-Huan Kuo ◽

Yung-Hsiang Chen

Keyword(s):

Neural Network ◽

Deep Neural Network ◽

Absolute Error ◽

Mean Absolute Percentage Error ◽

The Other ◽

Percentage Error ◽

Model Assessment ◽

Absolute Percentage Error ◽

Spatiotemporal Feature ◽

The Mean

AbstractThe coronavirus disease 2019 pandemic continues as of March 26 and spread to Europe on approximately February 24. A report from April 29 revealed 1.26 million confirmed cases and 125 928 deaths in Europe. This study proposed a novel deep neural network framework, COVID-19Net, which parallelly combines a convolutional neural network (CNN) and bidirectional gated recurrent units (GRUs). Three European countries with severe outbreaks were studied—Germany, Italy, and Spain—to extract spatiotemporal feature and predict the number of confirmed cases. The prediction results acquired from COVID-19Net were compared to those obtained using a CNN, GRU, and CNN-GRU. The mean absolute error, mean absolute percentage error, and root mean square error, which are commonly used model assessment indices, were used to compare the accuracy of the models. The results verified that COVID-19Net was notably more accurate than the other models. The mean absolute percentage error generated by COVID-19Net was 1.447 for Germany, 1.801 for Italy, and 2.828 for Spain, which were considerably lower than those of the other models. This indicated that the proposed framework can accurately predict the accumulated number of confirmed cases in the three countries and serve as a crucial reference for devising public health strategies.

Download Full-text

A Novel Neural Network Model for Blood Pressure Estimation Using Photoplethesmography without Electrocardiogram

Journal of Healthcare Engineering ◽

10.1155/2018/7804243 ◽

2018 ◽

Vol 2018 ◽

pp. 1-9 ◽

Cited By ~ 15

Author(s):

Ludi Wang ◽

Wei Zhou ◽

Ying Xing ◽

Xiaoguang Zhou

Keyword(s):

Neural Network ◽

Blood Pressure ◽

Mean Absolute Error ◽

Wearable Sensors ◽

Absolute Error ◽

Noninvasive Estimation ◽

Treatment Of Hypertension ◽

The Mean ◽

Artificial Neural Network Ann ◽

Blood Pressure Estimation

The prevention, evaluation, and treatment of hypertension have attracted increasing attention in recent years. As photoplethysmography (PPG) technology has been widely applied to wearable sensors, the noninvasive estimation of blood pressure (BP) using the PPG method has received considerable interest. In this paper, a method for estimating systolic and diastolic BP based only on a PPG signal is developed. The multitaper method (MTM) is used for feature extraction, and an artificial neural network (ANN) is used for estimation. Compared with previous approaches, the proposed method obtains better accuracy; the mean absolute error is 4.02 ± 2.79 mmHg for systolic BP and 2.27 ± 1.82 mmHg for diastolic BP.

Download Full-text

A method for predicting the capacity of a lithium-ion battery based on deep neural network

10.31224/osf.io/bzwkp ◽

2021 ◽

Author(s):

JamesChan

Keyword(s):

Neural Network ◽

Deep Neural Network ◽

Lithium Ion ◽

Chemical Conversion ◽

Absolute Error ◽

Production Costs ◽

Percentage Error ◽

Average Percentage ◽

True Value ◽

Battery Capacity

This paper proposes a solution to predict the capacity of the lithium-ion battery's capacity division process using deep learning methods. This solution extracts the physical observation records of part of the process steps from the chemical conversion and volumetric processes as features, and trains a Deep Neural Network (DNN) to achieve accurate prediction of battery capacity. According to the test, the average percentage absolute error (Mean Absolute Percentage Error, MAPE) of the battery capacity predicted by this model is only 0.78% compared with the true value. Combining this model with the production line can greatly reduce production time and energy consumption, and reduce battery production costs.

Download Full-text

Forecasting Malaysian overnight islamic interbank rate using the Box-Jenkins model

Data Analytics and Applied Mathematics (DAAM) ◽

10.15282/daam.v2i1.6837 ◽

2021 ◽

Vol 2 (1) ◽

pp. 38-51

Author(s):

N.S.M. Radzi ◽

S.R. Yaziz

Keyword(s):

Monetary Policy ◽

Mean Absolute Error ◽

Absolute Error ◽

Prediction Intervals ◽

Percentage Error ◽

Mean Square ◽

Islamic Banks ◽

Financial Instruments ◽

Policy Makers ◽

Forecasting Performance

Modelling the overnight Islamic interbank rate (IIR) is imperative to define the IIR performance as it would help the Islamic banks to adjust its costs of funding effectively and facilitate the policy makers to regulate a comprehensive monetary policy in Malaysia. The IIR framework which has been regulated by Bank Negara Malaysia under dual banking and financial system has always been overlooked in most previous studies in modelling the financial instruments rates. Therefore, it is vital to select the appropriate model as it resembles with the features of the IIR. The study assesses the forecasting performance of overnight IIR using the Box-Jenkins model. The suggested Box-Jenkins model has been applied to the Malaysian overnight IIR (in percentage) from 02/01/2001 to 31/12/2020. The empirical results determine that ARIMA (0,1,1) is the most appropriate model in forecasting overnight IIR as the model provides the smallest Mean Absolute Error (MAE), Root Mean Square Error (RMSE) and Mean Absolute Percentage Error (MAPE). In multistep ahead forecasting, it can be summarised that ARIMA (0,1,1) model is able to trail the actual data trend of daily Malaysian overnight IIR up to 5-day ahead within 95% prediction intervals.

Download Full-text

Development and Evaluation of the Combined Machine Learning Models for the Prediction of Dam Inflow

Water ◽

10.3390/w12102927 ◽

2020 ◽

Vol 12 (10) ◽

pp. 2927

Author(s):

Jiyeong Hong ◽

Seoro Lee ◽

Joo Hyun Bae ◽

Jimin Lee ◽

Woon Ji Park ◽

...

Keyword(s):

Neural Network ◽

Machine Learning ◽

Random Forest ◽

Multilayer Perceptron ◽

Short Term Memory ◽

Learning Algorithms ◽

Absolute Error ◽

Machine Learning Algorithms ◽

Gradient Boosting ◽

Dam Inflow

Predicting dam inflow is necessary for effective water management. This study created machine learning algorithms to predict the amount of inflow into the Soyang River Dam in South Korea, using weather and dam inflow data for 40 years. A total of six algorithms were used, as follows: decision tree (DT), multilayer perceptron (MLP), random forest (RF), gradient boosting (GB), recurrent neural network–long short-term memory (RNN–LSTM), and convolutional neural network–LSTM (CNN–LSTM). Among these models, the multilayer perceptron model showed the best results in predicting dam inflow, with the Nash–Sutcliffe efficiency (NSE) value of 0.812, root mean squared errors (RMSE) of 77.218 m3/s, mean absolute error (MAE) of 29.034 m3/s, correlation coefficient (R) of 0.924, and determination coefficient (R2) of 0.817. However, when the amount of dam inflow is below 100 m3/s, the ensemble models (random forest and gradient boosting models) performed better than MLP for the prediction of dam inflow. Therefore, two combined machine learning (CombML) models (RF_MLP and GB_MLP) were developed for the prediction of the dam inflow using the ensemble methods (RF and GB) at precipitation below 16 mm, and the MLP at precipitation above 16 mm. The precipitation of 16 mm is the average daily precipitation at the inflow of 100 m3/s or more. The results show the accuracy verification results of NSE 0.857, RMSE 68.417 m3/s, MAE 18.063 m3/s, R 0.927, and R2 0.859 in RF_MLP, and NSE 0.829, RMSE 73.918 m3/s, MAE 18.093 m3/s, R 0.912, and R2 0.831 in GB_MLP, which infers that the combination of the models predicts the dam inflow the most accurately. CombML algorithms showed that it is possible to predict inflow through inflow learning, considering flow characteristics such as flow regimes, by combining several machine learning algorithms.

Download Full-text

A Comparison of Models for Estimating Solar Radiation from Sunshine Duration in Croatia

International Journal of Photoenergy ◽

10.1155/2020/9605950 ◽

2020 ◽

Vol 2020 ◽

pp. 1-14

Author(s):

Tihomir Betti ◽

Ivana Zulim ◽

Slavica Brkić ◽

Blanka Tuka

Keyword(s):

Mean Absolute Error ◽

Sunshine Duration ◽

Model Performance ◽

Absolute Error ◽

Mean Bias Error ◽

Percentage Error ◽

Bias Error ◽

Mean Square ◽

Using Data ◽

Statistical Indicators

The performance of seventeen sunshine-duration-based models has been assessed using data from seven meteorological stations in Croatia. Conventional statistical indicators are used as numerical indicators of the model performance: mean absolute percentage error (MAPE), mean bias error (MBE), mean absolute error (MAE), and root-mean-square error (RMSE). The ranking of the models was done using the combination of all these parameters, all having equal weights. The Rietveld model was found to perform the best overall, followed by Soler and Dogniaux-Lemoine monthly dependent models. For three best-performing models, new adjusted coefficients are calculated, and they are validated using separate dataset. Only the Dogniaux-Lemoine model performed better with adjusted coefficients, but across all analysed locations, the adjusted models showed improvement in reduced maximum percentage error.

Download Full-text

On Mean Absolute Error for Deep Neural Network Based Vector-to-Vector Regression

IEEE Signal Processing Letters ◽

10.1109/lsp.2020.3016837 ◽

2020 ◽

Vol 27 ◽

pp. 1485-1489 ◽

Cited By ~ 1

Author(s):

Jun Qi ◽

Jun Du ◽

Sabato Marco Siniscalchi ◽

Xiaoli Ma ◽

Chin-Hui Lee

Keyword(s):

Neural Network ◽

Deep Neural Network ◽

Mean Absolute Error ◽

Absolute Error

Download Full-text

Detection of Tea Export Potentiality from Bangladesh in the World Market: An Autoregressive Integrated Moving Average (ARIMA) Approach

Asian Journal of Agricultural Extension Economics & Sociology ◽

10.9734/ajaees/2020/v38i430333 ◽

2020 ◽

pp. 21-29

Author(s):

A. U. Noman ◽

S. Majumder ◽

M. F. Imam ◽

M. J. Hossain ◽

F. Elahi ◽

...

Keyword(s):

Mean Absolute Error ◽

Moving Average ◽

Arima Model ◽

Absolute Error ◽

World Market ◽

Information Criteria ◽

Coefficient Of Determination ◽

Percentage Error ◽

Autoregressive Integrated Moving Average ◽

The World

Export plays an important role in promoting economic growth and development. The study is conducted to make an efficient forecasting of tea export from Bangladesh for mitigating the risk of export in the world market. Forecasting has been done by fitting Box-Jenkins type autoregressive integrated moving average (ARIMA) model. The best ARIMA model is selected by comparing the criteria- coefficient of determination (R2), root mean square error (RMSE), mean absolute percentage error (MAPE), mean absolute error (MAE) and Bayesian information criteria (BIC). Among the Box-Jenkins ARIMA type models for tea export the ARIMA (1,1,3) model is the most appropriate one for forecasting and the forecast values in thousand kilogram for the year 2017-18, 2018-19, 2019-20, 2020-21 and 2021-22, are 1096.48, 812.83, 1122.02, 776.25 and 794.33 with upper limit 1819.70, 1348.96, 1862.09, 1288.25, 1318.26 and lower limit 660.69, 489.78, 676.08, 467.74, 478.63, respectively. So, the result of this model may be helpful for the policymaker to make an export development plan for the country.

Download Full-text