USA housing-price-prediction using regression This project focuses on predicting house prices in the USA using a machine learning model. The process involves thorough data exploration, preprocessing, and wrangling to ensure the dataset is suitable for training. Feature selection is performed using the forward method, and the model is trained using TensorFlow with a simple Neural Network architecture.
Key Steps Data Exploration:
Analyzed and understood the dataset to identify patterns, outliers, and potential relationships. Data Preprocessing:
Cleaned and prepared the data for modeling, handling missing values, outliers, and ensuring consistency. Feature Selection:
Utilized forward method feature selection techniques to choose the most relevant features for the prediction model. Model Training:
Implemented a Neural Network using TensorFlow for house price prediction. Model Evaluation:
Achieved an MSE (Mean Squared Error) of 2% on the training set and 7% on the testing set, demonstrating the model's effectiveness.