SJIF Impact Factor (2023): 8.574| ISI I.F. Value: 1.241| Journal DOI: 10.

36713/epra2016 ISSN: 2455-7838(Online)

EPRA International Journal of Research and Development (IJRD)
Volume: 8 | Issue: 10 | October 2023 - Peer Reviewed Journal

FLIGHT FARE PREDICTION USING MACHINE LEARNING

K.D.V.N.Vaishnavi1, L. Hima Bindu2, M. Satwika3, K. Udaya Lakshmi4, M. Harini5,

N. Ashok6
1,2,3,4,5
Students, Information Technology, Vasireddy Venkatadri Institute of technology, Guntur, India.
1
Department of Information Technology
1
Vasireddy Venkatadri Institute of Technology, Guntur, India.
6
Associate Professor, Department of Information Technology, Vasireddy Venkatadri Institute of Technology
Guntur, Andhra Pradesh, India.

Article DOI: https://doi.org/10.36713/epra14763

DOI No: 10.36713/epra14763

ABSTRACT
The Flight Fare Prediction System is a comprehensive solution aimed at accurately forecasting flight ticket prices, providing
travelers with valuable insights for better planning and decision-making. Nowadays, airline ticket prices can vary dynamically
for the same flight. From the customer's perspective, they want to save money, so I have proposed a model that predicts the
approximate ticket price. This system leverages machine learning algorithms and historical flight data to generate accurate fare
predictions. The system utilizes a vast dataset comprising historical flight fares, including factors such as travel dates,
destinations, airlines, departure times, and various other relevant variables. By analyzing this data using advanced machine
learning techniques, the system learns patterns and relationships, enabling it to make reliable predictions about future flight fares.
An ensemble of machine learning algorithms, including regression-based models like Random Forest, Gradient Boosting, and
Support Vector Regression, is employed to capture complex patterns and relationships within the data. This system will give people
an idea of the trends the prices follow and also provide the predicted value of the price, which they can check before booking flights
to save money. This kind of system or service can be provided to customers through flight booking companies to help
them book tickets.
KEYWORDS - Flight Fare Prediction, Machine Learning, Historical Flight data, Random Forest.

I. INTRODUCTION
Everyone knows that holidays always call for a much-needed vacation and planning the travel itinerary becomes a time-consuming
task. The commercial aviation business has grown tremendously and has become a regulated marketplace as a result of the
worldwide growth of the Internet and E-commerce. Hence, for Airline revenue management, different strategies like customer
profiling, financial marketing, and social factors are used for setting ticket fairs. When tickets are booked months in advance, airfares
are often reasonable, but when tickets are booked in a hurry, they are often higher. But, the number of days/hours until departure
isn’t the only factor that decides flight fare, there are numerous other factors as well. Customers find it quite difficult to obtain a
perfect and lowest ticket deal due to the aviation industry's complex pricing methodology. Machine Learning and Deep Learning-
based technologies and models have been created to overcome this challenge, and substantial research is also happening. This study
discusses a Machine Learning-based Flight Fare Prediction System that employs Random Forest Regression to predict airline ticket
pricing. Various features that influence prices are also studied along with the system's experimental analysis. Section II included a
literature review that looked at technical papers as well as some current models and systems. Differences in the features considered
are also mapped down, In Section III, the proposed system is described in detail along with the workflow and its features. In Section
IV, the results as well as various comparisons between findings are reported. In Section V, conclusions are provided. In Section VI,
prospective advancements for further research.

II. RELATED WORK

The paper begins with some broad information regarding machine learning, after which the authors further proceed to the
methodology. The methodology consists of four-phase process that influences flight prices, collection of data from flight fare
prediction dataset by MH, selection and evaluation of an accurate ML Regression model. Key to its success is the integrity of the
system where consumers accurately represent the segments for which prices have been differentially determined.

2023 EPRA IJRD | Journal DOI: https://doi.org/10.36713/epra2016 | https://eprajournals.com/ |245 |

 SJIF Impact Factor (2023): 8.574| ISI I.F. Value: 1.241| Journal DOI: 10.36713/epra2016 ISSN: 2455-7838(Online)
EPRA International Journal of Research and Development (IJRD)
Volume: 8 | Issue: 10 | October 2023 - Peer Reviewed Journal

[2] Today, airlines price tickets “as much as the customer and market will bear,” according to consultant and former airline planning
executive. Airlines also profile their customers to help them adjust prices.
[3] This often means placing passengers into one of two groups: leisure or business. And the way each group is priced is very
different. Most studies on airfare price prediction have focused on either the national level or a specific market. Research at the
market segment level, however, is still very limited. We define the term market segment as the market/airport pair between the flight
origin and the destination.

[4] The airline dataset included the following eight characteristics: departure and arrival times, type of airline, number of stopages,
source, destination and additional_information. The authors performed prediction using regression Machine Learning models that
including, LGBM Regressor, Random Forest Regression Tree, and Decision Tree Regressor. Random Forest Regression is a
supervised learning algorithm that uses ensemble learning method for regression. Ensemble learning method is a technique that
combines predictions from multiple machine learning algorithms to make a more accurate prediction than a single model.

[5] All in present a review of deep learning and social media data-based Airline ticket price prediction model. The authors introduce
the current airline ticket pricing situation with the factors that affect ticket prices. A Random Forest operates by constructing several
decision trees during training time and outputting the mean of the classes as the prediction of all the trees. A Random Forest
Regression model is powerful and accurate. It usually performs great on many problems, including features with non-linear
relationships.

[6] Random Forest Model is used for development since it outperforms other models such as LGBM Decision Tree Regressor and
Neural Network in terms of data performance with a R squared score of 0.91, this prediction framework has a good level of accuracy.

III. METHODOLOGY
The following steps are involved in the proposed architecture of our project.
3.1. Data Collection
The act of obtaining, acquiring, and combining the data that will be used to develop, test, and verify a machine learning model is
known as data collection in machine learning. This step plays crucial role in implementation. Here data is collected from flight fare
dataset which is imported from Kaggle. The dataset consists of both categorical data and numerical data. The categorical data
includes source, destination, type of airline, additional info and numerical data includes arrival and departure dates, number of Stops.
There are 11 columns (each represents a feature) and 10683 rows in this large dataset.

3.2. Data Preprocessing

Data preprocessing means nothing but cleaning data, which can be used for model training and testing. By this step we can make
our data useful for model training purpose. Data preprocessing involves cleaning, transforming, and preparing the data for data
analysis. The sub steps involved in the data preprocessing are:

Data Cleaning: In this step the null values are removed, missing values are removed and if any duplicates are present that are also
removed.

Feature selection and engineering: In this step the features of our model are extracted and all the relevant features are used for
model training. In dataset it contains date of journey, arrival date, departure date columns and all the numerical values are extracted
as Departure hour, departure minutes, arrival hour, arrival minutes, journey day, journey month. As dataset contains both categorical
and numerical features, by using 'On hot encoding' method for nominal categorical data and 'label encoding' for ordinal categorical
data was used to convert the categorical values to numerical values. The dataset consists of categorical variables like airline, source,
destination, route, total number of stops and additional info.

3.3 Data Splitting

This step involves splitting our data into two parts for training and testing purpose. For model training 80 percent of data was used
by using Random Forest regressor model was trained. The machine learning algorithms are:

LGBM Regressor
LGBM stands for Light Gradient Boosted Machine. It is a gradient boosting framework based on decision trees that can be used for
various machine learning tasks such as regression, classification and ranking. LGBM Regressor is a class in lightgbm package that
can be used to train and predict regression models.

2023 EPRA IJRD | Journal DOI: https://doi.org/10.36713/epra2016 | https://eprajournals.com/ |246 |

 SJIF Impact Factor (2023): 8.574| ISI I.F. Value: 1.241| Journal DOI: 10.36713/epra2016 ISSN: 2455-7838(Online)
EPRA International Journal of Research and Development (IJRD)
Volume: 8 | Issue: 10 | October 2023 - Peer Reviewed Journal

Decision Tree Regressor

Decision tree regressor is a class in Sklearn tree module that can be used to train and predict regression models. It is a decision tree-
based algorithm that recursively partitions the input data based on the values of the input features, forming a tree-like structure. It
initially chooses independent variable from dataset as decision nodes for decision making and then it divides the entire dataset into
sub-sections and when test data is passed to the model the output is decided by checking the data point belongs to the decision tree
will give output as the average value of all the datapoints in the sub-section.

Randon Forest Regressor

Random Forest regressor uses multiple decision trees to perform regression tasks. It is an example of ensemble learning. Random
forest is a Supervised Learning algorithm which uses ensemble learning approach for classification and regression. Decision trees
are sensitive to the specific data on which they are trained. If the training data is changed the resulting decision tree can be quite
different and in turn the predictions can be quite different. Also, Decision trees are computationally expensive to train, carry a big
risk of overfitting, and tend to find local optimal because they can’t go back after they have made a split to address these weaknesses,
we turn to Random Forest.

3.4 Model evaluation

This is an important step in our project, as it helps us to measure the performance and accuracy of our model. Test data is used for
model evaluation. Here, we employed Cross-validation for model evaluation. This method divides the data into k-subsets, called
folds. the model is trained on k-1 folds and tested on the remaining fold. this process is repeated k times, so that each fold is used
as a test set once. The average performance across all k-folds is reported as the final result. The metrics that are used for model
evaluation purpose are:

Root Mean Squared Error (RSME): It gives the root of the average squared difference between the actual values and the predicted
values for a regression problem.

Mean Absolute Error (MAE): It gives the absolute difference between the actual values and predicted values. The higher negative
mean values indicate the better performance of model.

R-Squared: This metric measures how well the regression model fits the data, by comparing it to a baseline model that always
predict the mean value. It shows how much variation in the data is explained by the model.

3.5 Model Architecture

The architecture is crafted with essential components to optimize prediction. Our proposed system analyses historical data to identify
patterns, seasonal trends and additional information that influence flight fares.

Fig.1: System Architecture

2023 EPRA IJRD | Journal DOI: https://doi.org/10.36713/epra2016 | https://eprajournals.com/ |247 |
 SJIF Impact Factor (2023): 8.574| ISI I.F. Value: 1.241| Journal DOI: 10.36713/epra2016 ISSN: 2455-7838(Online)
EPRA International Journal of Research and Development (IJRD)
Volume: 8 | Issue: 10 | October 2023 - Peer Reviewed Journal

IV. IMPLEMENTATION AND RESULTS

1. Importing the required libraries.

2. Read the dataset:

Our dataset format might be in .csv, excel files, .txt, .json, etc. We can read the dataset with the help of pandas.

3. Data preprocessing:
The df.isnull() method is used to verify that no values are present. We employ the sum () function to add up those null
values. Two null values were discovered in our dataset, we discovered. We thus start by investigating the data.

4. Using a Heat Map to check the correlation:

Here we are using a heat map to check the correlation in this instance. Using different colour combinations, it displays the
data as 2-D coloured maps. Instead of numbers, it will be plotted on both axes to describe the relationship variables.

2023 EPRA IJRD | Journal DOI: https://doi.org/10.36713/epra2016 | https://eprajournals.com/ |248 |

 SJIF Impact Factor (2023): 8.574| ISI I.F. Value: 1.241| Journal DOI: 10.36713/epra2016 ISSN: 2455-7838(Online)
EPRA International Journal of Research and Development (IJRD)
Volume: 8 | Issue: 10 | October 2023 - Peer Reviewed Journal

5. By comparing all the models (LGBM Regressor, Decision Tree Regressor, Random Forest Regressor), we can conclude
that Random Forest Regressor performs the best.

Fig 2: Flight Price Test and Predicted Data

2023 EPRA IJRD | Journal DOI: https://doi.org/10.36713/epra2016 | https://eprajournals.com/ |249 |

 SJIF Impact Factor (2023): 8.574| ISI I.F. Value: 1.241| Journal DOI: 10.36713/epra2016 ISSN: 2455-7838(Online)
EPRA International Journal of Research and Development (IJRD)
Volume: 8 | Issue: 10 | October 2023 - Peer Reviewed Journal

Fig.3: Home Page

Figure 3 By using flask the machine learning model wasy deployed. Flask is a python web frame work that allows you to build
lightweight and flexible web applications quickly and easily. The web page collects information and it predicts the fare of ticket. It
collects information like departure date, arrival date, source, destination, number of stops, type of airline and additional
information(meal included or not, no info, change airports, etc).

V. FUTURE SCOPE
➢ Optimal date recommendation: It means suggesting the date to users on which date the flight prices will be minimum.
➢ Real-Time Updates: Dealing with real-time data for dynamic pricing adjustments based on factors like weather, demand,
and airline policies.
➢ Integration: Partnering with airlines, travel agencies, and online booking platforms to provide pricing as a value-added
service.

VI. CONCLUSION
In conclusion, the main aim of our project flight fare prediction using machine learning is to predict the prices. we have created a
User Interface for the entire process which includes arrival date, departure date, source, destination, etc. Our flight fare prediction
project using machine learning has successfully produced a reliable and user-friendly system. We collected, preprocessed, and
extracted features from flight fare data, trained a robust random forest model and evaluated its performance. This web application
we developed empowers travelers to make informed decisions by predicting flight prices based on their input.

VII. REFERENCES
1. K. Tziridis, Th. Kalampokas, G. A. Papakostas, “Airfare Prices Prediction Using Machine Learning Techniques”, 25th European Signal
Processing Conference (EUSIPCO), IEEE, October 26, 2017.
2. N. Brown and J. Taylor, Air Fare: Stories, Poems & Essays on Flight. Sarabande Books, 2004.
3. J. C. Lok, Prediction Factors Influence Airline Fuel Price Changing Reasons. 2018.
4. A. Kakaraparthi and V. Karthick, “A secure and cost-effective platform for employee management system using lightweight standalone
framework over Diffie Hellman’s key exchange algorithm,” ECS Trans., vol. 107, no. 1, pp. 13663–13674, Apr.2022.
5. Wikipedia, “MeanSquarederror”, Available:https://en.wikipedia.org/wiki/Mean_squared_error.
6. G. Ataman and S. Kahraman, “Stock Market Prediction in Brics Countries Using Linear Regression and Artificial Neural Network
Hybrid Models,” The Singapore Economic Review. pp. 1–19, 2021. doi: 10.1142/s0217590821500521.
7. B. Panwar, G. Dhuriya, P. Johri, S. S. Yadav, and N. Gaur, “Stock Market Prediction Using Linear Regression and SVM,” 2021
International Conference on Advance Computing and Innovative Technologies in Engineering (ICACITE). 2021. doi:
10.1109/icacite51222.2021.9404733.
8. P. Purey and A. Patidar, “Stock Market Close Price Prediction Using Neural Network and Regression Analysis,” International Journal
of Computer Sciences and Engineering, vol. 6, no. 8. pp. 266–271, 2018. doi: 10.26438/ijcse/v6i8.266271.
9. William Groves and Maria Gini "An agent for optimizing airline ticket purchasing" in proceedings of the 2013 international conference
on autonomous agents and multi-agent systems.
10. https://towardsdatascience.com/machine-learning-basics-decisiontree-regression-1d73ea003fda article on decision tree regression.
11. www.keboola.com/blog/random-forest-regression article on random forest.

2023 EPRA IJRD | Journal DOI: https://doi.org/10.36713/epra2016 | https://eprajournals.com/ |250 |