Recurrent Neural Networks and its Applications in Time Series Data

Le résumé fourni par la source

Recurrent Neural Networks (RNNs) have been extensively embraced for sequencing analysis and sequential data modelling, especially time series data. This paper delves into the principles and applications of the RNNs concentrically time series analysis. The advantages of time series RNNs, including long short-term memory (LSTM) and Gated recurrent units (GRU), are discussed in detail. Different RNN models have been implemented on both synthetic and real-look time series datasets to assess their performance. The results clearly show RNNs outperformed traditional forecasting and pattern recognition methods. Lastly, the issues related to RNN training, namely, vanishing gradients and overfitting, are briefly addressed, and potential improvements for RNNs are outlined.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.