Authors :
Pranjal Sharma; Prashant Singh; Plash Upreti
Volume/Issue :
Volume 6 - 2021, Issue 7 - July
Google Scholar :
http://bitly.ws/9nMw
Scribd :
https://bit.ly/3fdSaAq
Abstract :
Our Research paper proposes a model of
travel time based prediction using regression techniques.
The objective of our model is to predict the accurate trip
duration of a taxi from one of the pickup location to
another dropoff location. In today’s fast-paced world,
where everyone is short of time and is always in a hurry,
everyone wants to know the exact duration to reach
his/her destination to carry ahead of their plans. So, for
their serenity, we already have million dollar startups
such as Uber and Ola where we can track our trip
duration. As a result of this, we proposed a technique in
which every cab service provider can give exact trip
duration to their customers taking into consideration the
factors such as traffic, time and day of pickup. So, in our
methodology, we propose a method to make predictions
of trip duration, in which we have used several
algorithms, tune the corresponding parameters of the
algorithm by analyzing each parameter against RMSE
and predict the trip duration. To make our prediction
we used RandomForest Regressor, LinearSVR and
LinearRegression. We improved the accuracy by tuning
hyperparameters and RandomForest gave the best
accuracy.
We also analyzed several data mining techniques to
handle missing data, remove redundancy and resolve
data conflicts. We used the NYC Limousine OpenData,
and the travel details of the month of January in the year
2015 to carry ahead with feature extraction and
prediction.
Our Research paper proposes a model of
travel time based prediction using regression techniques.
The objective of our model is to predict the accurate trip
duration of a taxi from one of the pickup location to
another dropoff location. In today’s fast-paced world,
where everyone is short of time and is always in a hurry,
everyone wants to know the exact duration to reach
his/her destination to carry ahead of their plans. So, for
their serenity, we already have million dollar startups
such as Uber and Ola where we can track our trip
duration. As a result of this, we proposed a technique in
which every cab service provider can give exact trip
duration to their customers taking into consideration the
factors such as traffic, time and day of pickup. So, in our
methodology, we propose a method to make predictions
of trip duration, in which we have used several
algorithms, tune the corresponding parameters of the
algorithm by analyzing each parameter against RMSE
and predict the trip duration. To make our prediction
we used RandomForest Regressor, LinearSVR and
LinearRegression. We improved the accuracy by tuning
hyperparameters and RandomForest gave the best
accuracy.
We also analyzed several data mining techniques to
handle missing data, remove redundancy and resolve
data conflicts. We used the NYC Limousine OpenData,
and the travel details of the month of January in the year
2015 to carry ahead with feature extraction and
prediction.
