Flight-Fare-Prediction

Objective

The objective of this article is to predict flight prices given the various parameters. Data used in this article is publicly available at Kaggle. This will be a regression problem since the target or dependent variable is the price (continuous numeric value).

Introduction

Airlines use complex algorithms to calculate flight prices based on the many factors that exist at the moment. To predict flight fares, these strategies consider financial, marketing, and numerous societal elements.

The number of people who fly has dramatically increased in recent years. Pricing alter dynamically owing to many variables, making it difficult for airlines to maintain prices. As a result, we will attempt to solve this problem using machine learning. This can assist airlines in determining what rates they can keep. Customers can also use it to forecast future airline prices and plan their trip appropriately.

Approach

• Data was used from Kaggle which is a freely available platform for data scientists and machine learning enthusiasts. We are using jupyter-notebook to run Flight Price Prediction task.
• Data Analysis
• Data Preparation
• Model Building
• Predicting the model on test data

Usage

• Just run jupyter notebook in terminal and it will run in your browser.

  Install Jupyter here i've you haven't.

Modules needed:

- NumPy
- Pandas
- Scikit-Learn &
- seaborn

Models with their Accuracy of Prediction

• The machine learnig model accuracy is around 83%
• Hypertuning the model GridSearch CV is a technique used to validate the model with different parameter combinations, by creating a grid of parameters and trying all the combinations to compare which combination gave the best results. We apply grid search on our model – accuracy 87%

Steps to run this application in your system

• Clone or download the repo.
• Open command prompt in the downloaded folder.

Dataset Link

https://www.kaggle.com/nikhilmittal/flight-fare-prediction-mh