Sales Prediction Task Overview This project was completed as part of my internship at Codsoft. The objective of the task was to develop a predictive model to forecast sales based on advertising expenditures across different channels, namely TV, Radio, and Newspaper.
Dataset The dataset provided for this task contained information on advertising budgets for TV, Radio, and Newspaper, as well as corresponding sales figures. It served as the foundation for building and evaluating the predictive model.
Approach The task was approached as a supervised regression problem. Various machine learning algorithms were explored to predict sales, including linear regression, random forest, gradient boosting, and ensemble methods such as bagging.
Key Steps Data Exploration: Conducted initial exploration to understand the structure, distribution, and relationships within the dataset. Data Preprocessing: Handled any missing values, outliers, or other anomalies present in the dataset. Performed feature engineering if necessary.In this dataset,there is only 2 outliers with no null values,so work become simplified. Model Selection: Evaluated different regression algorithms and ensemble methods to identify the most suitable model for predicting sales. Model Training: Trained the selected model(s) using the training dataset, tuning hyperparameters as needed. Model Evaluation: Assessed the performance of the trained model(s) using relevant evaluation metrics such as Mean Absolute Error (MAE), Root Mean Squared Error and R^score. Cross-Validation: Employed cross-validation techniques to validate the robustness and generalization ability of the model(s).