SETalyze

Overview

SETalyze is a project aimed at predicting daily stock prices for Thailand's SET50 index using a combination of news, fundamental, and technical data. By employing advanced machine learning, SETalyze can make precise time series predictions, benefiting both traders and investors.

Objective

• Accuracy: Ensure high accuracy in predicting daily stock prices for trading and investment decisions.
• Directional Insight: Ascertain the probable directional movement (rise or fall) of a stock price.

Scope

• Focuses on predicting the stock's closing prices.
• Restricted to the SET50 index of the Thai stock market.

Getting Started

Prerequisites

• Python 3.x
• Pandas
• Darts
• Scikit-learn
• Os
• Numpy

You can install the required packages using pip:

pip install pandas darts scikit-learn os numpy

Execution

To run SETalyze:

python prediction.py

Upon successful execution, predictions will be organized and stored in the result/ directory as Prediction.csv.

Methodology

1. Data Preprocessing: The system meticulously cleans the data, ensuring all irrelevant or incomplete data points are excluded. Subsequently, it scales the dataset and identifies key components to be retained for analysis.

2. Machine Learning: SETalyze harnesses the power of the BlockRNNModel from the Darts library to deliver accurate time series predictions.

3. Post-Processing & Output: After the prediction phase, SETalyze structures the outputs, ensuring they're intuitive and ready for any subsequent financial analyses. The results are then saved in a well-formatted CSV file.