Price Forecasting for Agricultural Products Based on Long Short Term Memory Model

Autor: Lin,Dai-Wei, 林大為
Rok vydání: 2018
Druh dokumentu: 學位論文 ; thesis
Popis: 106
This study aims to compare the predictive power of LSTM memory models with the traditional time series models. This study takes into account the impact of price and time on the model and aims to improve the prediction accuracy by changing the variables in the LSTM memory models. In the proposed study, the wholesale price of tomatoes was used as an example, and the price information of the Taichung market from 2003 to 2017 was considered. The results show that the LSTM models have an improved predictive ability. According to the predictive accuracy, the predictive ability of the model can be enhanced by referring to the highly correlated fruit and vegetable price. The model is capable of making relatively accurate predictions about one year in advance.
Databáze: Networked Digital Library of Theses & Dissertations