README.md

Sales Prediction Documentation

Introduction

Sales prediction is crucial for businesses as it aids in decision-making, resource allocation, and revenue forecasting. Leveraging machine learning techniques offers insights into the relationship between advertising expenses and sales, enabling effective marketing strategy optimization. This documentation provides an overview of a Python code implemented using Scikit-learn for sales prediction, covering data preprocessing, exploratory data analysis (EDA), model training, prediction, and evaluation.

Importing Libraries

The code initiates by importing essential libraries:

• Pandas: Data manipulation.
• NumPy: Numerical operations.
• Seaborn and Matplotlib: Data visualization.
• Scikit-learn: Machine learning algorithms.

Loading Dataset

The dataset provided by AFAME TECHNOLOGIES is loaded using Pandas' read_csv function, containing 200 rows of sales data with features including advertising expenses.

Pre-processing Data

Data pre-processing includes:

1. Data Augmentation: Adding random noise to increase data counts.
2. Combining Data: Combining augmented and original datasets for model training.

Exploratory Data Analysis (EDA)

Key EDA steps:

1. Data Description: Summary statistics for insights into distribution and variability.
2. Correlation Analysis: Visualizing relationships between features and the target variable.

Model Training, Prediction, and Evaluation

Multiple regression algorithms are employed:

1. Linear Regression: Establishing a linear relationship.
2. Polynomial Regression: Capturing nonlinear relationships.
3. Gradient Boosting Regressor: Utilizing ensemble methods.
4. Support Vector Machine (SVM): Predicting sales using SVM.
5. Neural Network: Constructing a feedforward neural network with TensorFlow-Keras.

Author Information

By Abhinav Mishra
Email: abhinavmishra@tuta.io