In this training, you will build a time-series data prediction model using NTT stock price data. We will start with exploratory data analysis (EDA) of the data and ultimately evaluate the accuracy of the model. We will use machine learning to predict stock prices and verify their accuracy, aiming for practical application.
- Data understanding and EDA
- Overview: Using NTT stock price data, confirm basic statistics, detect time series trends and seasonality, and confirm outliers.
- Goal: Understand data trends and extract issues for predictive models.
- Data Preprocessing and Feature Engineering
- Overview: Perform preprocessing such as missing value handling, normalization, and scaling to create features that are useful for stock price prediction.
- Objective: Prepare data to improve the accuracy of the model.
- Model Selection and Training
- Overview: Select and implement an appropriate predictive model. Examples include ARIMA models and recurrent neural networks such as LSTM.
- Objective: Build a model that can maximize the accuracy of stock price prediction.
- Model Evaluation and Results Analysis
- Overview: Verify the predictive accuracy of the model and calculate evaluation indicators.
- Objective: Objectively evaluate the effectiveness of the model and analyze the results.
- Consider improvement measures and retrain the model
- Overview: Find areas for improvement in the model and retrain it to reflect those improvements.
- Objective: Build a more accurate prediction model.
- Summary of results and preparation of presentation materials
- Overview: Summarize the results of the training as materials and report the model-building process and results in a presentation format.
- Objective: Develop the skills to communicate technical results in an easy-to-understand manner.