A Machine Learning application of using Neural Network model to predict house sale price

The first objective is to develop a predictive model, for this project I developed a neural network using tensorflow and keras adam optimizer. The second objective to improve real estate investment decisions using this model and informed decisions about future house investments . The third objective is assisting the buyers and sellers to establish a fair market and avoid any kind of manipulations and maintain transparency. The fourth and final objective is to familiarize graduates like us with application of the powerful tools and techniques of machine learning in tackling real world problems

In the current business world, nothing gets left behind without the involvement of technology, particularly the applications of Machine learning, it can be medical, education, government almost everywhere. And there is no shortage of the historical data that forms the basis for building and training any kind of model. Data driven decision making is one of the successful strategies across various industries around the globe making it valid proof decision. It is crucial to understand just building the model to predict something is not enough but the predictions should be accurate enough that provide better efficiency and help in improving the productivity.

The tools and techniques used for this analysis were pandas which is a python library for data manipulation and analysis, in order to load datasets and use them for model training, the scikit-learn used for splitting of the datasets into training and validation, Matplotlib and seaborn were used for visualizations and graphs to explore datasets or results. Keras (with TensorFlow backend): Used for building and training neural network models. Keras offers a user-friendly interface to create complex models with ease. The initial step involved loading the dataset and performing necessary preprocessing tasks. This included dropping irrelevant columns such as 'Id', handling missing values, and separating features from the target variable. Standardizing the features is a critical step in neural network modeling. The StandardScaler from Scikit-learn was used to scale the features

The scatter plot provides a visualization of the comparison of the predicted sale prices to the actual sale prices. The close clustering around the diagonal line shows that the model's predictions are reasonably accurate showing a strong correlation.