Multilayer Perceptron (MLP) and Autoregressive Integrated Moving Average (ARIMA) Models in Multivariate Input Time Series Data: Solar Irradiance Forecasting

Munandar, Devi

doi:10.18517/ijaseit.9.1.6426

Cited by 4 publications

(5 citation statements)

References 13 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…The MLP in this work consists of three layers, including input, hidden, and output layers. MLP layers are fully connected to each other and run mathematically as described in (19) [24] where Nt denotes the response variable in the output layer, which is also the predicted value of nonlinear components; εARIMA,t−i (i = 1, 2, . .…”

Section: Arima-mlp Modelmentioning

confidence: 99%

“…We define a cumulative distance Dist E , which can be expressed by (24). Dist E is the sum of the current local distance d E (S i , Ŝj ) and the distance from the current element to the nearest element.…”

Section: Dtw Algorithmmentioning

confidence: 99%

“…, (24) where Dist E (i, j) denotes the cost distance matric. And the initial condition is Dist E (1, 1) = d E X 1 , Y 1 .…”

Section: Dtw Algorithmmentioning

confidence: 99%

See 2 more Smart Citations

A new GNSS outlier mitigation method for GNSS/INS integrated system

Guo

Zhou

Zhao

et al. 2023

Meas. Sci. Technol.

View full text Add to dashboard Cite

High-precision positioning with Global Navigation Satellite Systems (GNSS) remains a significant challenge in urban environments, due to the outliers caused by the insufficient number of accessible satellites and environmental interference. A GNSS outlier mitigation algorithm with effective Fault Detection and Exclusion (FDE) is required for high-precision positioning. The traditional methods are designed to deal with zero-mean noise in GNSS, which leads to instabilities under biased measurements. Considering that GNSS data are typical time series data, a dynamic FDE scheme is constructed by combining a prediction-model-based method and a dissimilarity-based method. First, a hybrid prediction model which combines Autoregressive Integrated Moving Average (ARIMA) model and Multilayer Perceptron (MLP) model is proposed to provide pseudo-GNSS series by predicting the vehicle’s location for several future steps. Then, a dissimilarity-based method of Dynamic Time Warping (DTW) measure is utilized to analyze the pairwise dis-similarity between the pseudo-GNSS series and the received GNSS series. The performance of the different models in forecasting is evaluated, and the results show that the positioning accuracy is significantly improved by applying the ARIMA-MLP. The effectiveness of the proposed FDE method is verified through simulation experiments and real experiments based on a typical urban canyon public dataset collected in Tokyo.

show abstract

Section: Arima-mlp Modelmentioning

confidence: 99%

Section: Dtw Algorithmmentioning

confidence: 99%

See 1 more Smart Citation

A new GNSS outlier mitigation method for GNSS/INS integrated system

Guo

Zhou

Zhao

et al. 2023

Meas. Sci. Technol.

View full text Add to dashboard Cite

show abstract

“…A partir dessa informac ¸ão, descartou-se a possibilidade de usar modelos estatísticos no escopo do trabalho, como foi feito por Cedismondi [6], pois viu-se que o comportamento das taxas requereria modelos com maior capacidade de generalizac ¸ão, como redes neurais. O perceptron multi-camadas é considerado como uma das redes neurais mais simples e será utilizada neste trabalho como referência basal (baseline) para os demais métodos [11].…”

Section: Análise Do Dataset Consolidadounclassified

Modelos de Previsão das Taxas de Juros dos Títulos IPCA do Tesouro Direto

Oliveira¹,

Gomes²,

Santos³

et al. 2021

Anais Do 15. Congresso Brasileiro De Inteligência Computacional

View full text Add to dashboard Cite

Lançada pelo Governo Federal brasileiro no final de 2002, a plataforma Tesouro Direto possibilita a compra de títulos públicos por pessoas físicas, caracterizando-o como um investimento de renda fixa seguro e mais rentável que a poupança. Considerando as variações que as taxas de juros desses títulos sofrem ao longo do tempo, existe a possibilidade de se obter lucros acima do acordado inicialmente com a compra/venda no momento certo. Assim, este trabalho propõe a utilização de modelos de aprendizado de máquina, mais precisamente das redes neurais MLP, CNN e LSTM, na tentativa de se criar um modelo de previsão das taxas de juros de quatro títulos pós-fixados do Tesouro Direto atrelados ao IPCA. Para isso, foram analisadas e definidas algumas features (características) da economia brasileira que poderiam impactar mais diretamente nas taxas do Tesouro Direto. Simulações computacionais foram realizadas com variações de presença dessas características nos conjuntos de dados para que o melhor modelo de previsão fosse identificado. Ao final, concluiu-se que a CNN apresenta um melhor resultado geral, mesmo se mostrando mais sensível à remoção de features. Em contrapartida, a MLP apresenta erros quase constantes, independentemente da variável que está sendo removida. Além disso, descobriu-se que o CDI, importante métrica de investimentos, impacta de forma negativa a capacidade de previsão dos modelos, devido a sua proximidade de valor com a taxa Selic.

show abstract

“…Some prominent methods include clustering and classification techniques with hourly clearness index as the input [18], a simple regression model using the clearness index and relative sunshine [19], and a multivariate adaptive regression spline (MARS) technique with ten weather parameters as inputs [20]. The most recent notable methods include the use of a modification of empirical model [21], a hybrid of regression [22], [23], and autoregressive integrated moving average (ARIMA) [24]. Additionally, machine learning algorithms, particularly artificial neural networks (ANN), are widely used to predict hourly solar radiation.…”

Section: Introductionmentioning

confidence: 99%

Estimation of Hourly Solar Radiations on Horizontal Surface from Daily Average Solar Radiations Using Artificial Neural Network

Kurniawan¹,

Shintaku²

2022

Int. J. Adv. Sci. Eng. Inf. Technol.

View full text Add to dashboard Cite

Hourly solar radiation information is necessary to increase the effectiveness of photovoltaic (PV) energy systems. However, compared with daily average solar radiation, the measurement of hourly solar radiation is much less available. In this study, a multilayer perceptron artificial neural network based on daily average solar radiation is proposed to estimate the amount of hourly solar radiation. The proposed method also relies on the location of the PV system, the hour when the estimate is needed, and the month when the estimate is needed. Two separate networks were built to estimate hourly direct solar radiation and hourly diffuse solar radiation, the summation of which is the estimated value of global hourly solar radiation. The networks were trained using three years of data from 2016 to 2018 from five locations around Japan. The data were taken from the website of the Japan Meteorological Agency (JMA). The number of hidden neurons for each network was determined by comparing the regression value obtained during the training process. The proposed method was validated by comparing the estimated value with the actual measured solar radiation value for each month in 2019 for the same five locations used for training the networks. The reliability of the proposed method was confirmed with the minimum R 2 of 0.951 for the estimation of hourly direct solar radiation and 0.983 for hourly diffuse solar radiation. Further improvement in the accuracy of estimating the daily peak radiation value is considered to improve the total accuracy of the estimation model in the future.

show abstract

Multilayer Perceptron (MLP) and Autoregressive Integrated Moving Average (ARIMA) Models in Multivariate Input Time Series Data: Solar Irradiance Forecasting

Cited by 4 publications

References 13 publications

A new GNSS outlier mitigation method for GNSS/INS integrated system

A new GNSS outlier mitigation method for GNSS/INS integrated system

Modelos de Previsão das Taxas de Juros dos Títulos IPCA do Tesouro Direto

Estimation of Hourly Solar Radiations on Horizontal Surface from Daily Average Solar Radiations Using Artificial Neural Network

Contact Info

Product

Resources

About