Open AccessDeep Learning Approach in Predicting Property and Real Estate Indices
Author(s) -
Seng Hansun
Publication year - 2022
Publication title -
international journal of advances in soft computing and its applications
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.15
H-Index - 18
eISSN - 2710-1274
pISSN - 2074-8523
DOI - 10.15849/ijasca.220328.05
Subject(s) - real estate , deep learning , artificial intelligence , stock (firearms) , artificial neural network , computer science , stock market , property (philosophy) , econometrics , machine learning , economics , finance , engineering , geography , mechanical engineering , philosophy , context (archaeology) , archaeology , epistemology
The real estate market is one of the most impacted sectors from the Corona Virus Disease 2019 (COVID-19) pandemic that happened in early 2020 globally. Here, we tried to apply an extension of the Long Short-Term Memory (LSTM) deep learning method, known as the Bidirectional LSTM (Bi-LSTM) networks for stock price prediction. Our focus is on six stocks that were included in the LiQuid45 (LQ45) property and real estate sectors. A simple three-layers Bi-LSTM network is proposed for predicting the stocks’ closing prices. We found that the prediction results fall in the reasonable prediction category, except for Pembangunan Perumahan Tbk (PTPP). Bumi Serpong Damai Tbk (BSDE) got the highest accuracy result with more than 90% score, while PTPP got the lowest score with less than 8% score. The proposed Bi-LSTM network could provide a baseline result for developing a good trading strategy. Keywords: Bi-LSTM networks, deep learning, LQ45, property and real estate, stock price prediction.