|
Title:
|
REGIONAL GROSS DOMESTIC PRODUCT PREDICTION
USING TWITTER DEEP LEARNING REPRESENTATIONS |
|
Author(s):
|
Javier Ortega-Bastida, Antonio-Javier Gallego, Juan Ramón Rico-Juan and Pedro Albarrán |
|
ISBN:
|
978-989-8704-24-5 |
|
Editors:
|
Hans Weghorn |
|
Year:
|
2020 |
|
Edition:
|
Single |
|
Keywords:
|
Machine Learning, Neural Networks, Knowledge-Based Economy |
|
Type:
|
Full |
|
First Page:
|
89 |
|
Last Page:
|
96 |
|
Language:
|
English |
|
Cover:
|
|
|
Full Contents:
|
click to dowload 
|
|
Paper Abstract:
|
This work presents a method to predict the regional Gross Domestic Product (GDP) using the textual information stored
in tweets. In particular, we propose the use of a hybrid autoencoder to predict the GDP of the Valencian Community
(Spain) using the tweets written by the most influential economists, politicians, newspapers, and institutions in the region.
The proposed method uses an autoencoder that is trained to simultaneously minimize the reconstruction error of the
textual information of tweets along with their corresponding GDP value. In this way, we ensure that the latent space of
the autoencoder represents the important characteristics of tweets that relate them to the GDP. This system allows to
obtain a representation of new tweets that is related to a GDP prediction and also to analyze them to create clusters that
represent the agreement or disagreement of the opinions. The results show that the proposed approach captures the GDP
growth trend with a small margin of error. This method was also compared with other methods from the state of the art
showing a better result in prediction accuracy. |
|
|
|
|
|
|
