Estimation Of Arrival Time Of Taxi Based On Ensemble Learning

Nowadays,it is convenient to taking a taxi by using a Mobile phone application.How to offer the customers a good riding experience,which requires the shortest route and a correct estimated time of arrival(ETA)when the customer inputs the starting point and destination,is a good research target.A reasonable estimation of the trip duration is also very helpful for the taxi company to make efficient dispatch of its drivers and avoid the unnecessary waste.Machine learning has an irreplaceable effect on many fields this days.it brings us much convenience and even change our life style.this is a data era.We produce tons of data everyday.Machine learning can get valuable information on the basis of data modeling.And make decisions and predictions for us.Recently,ensemble learning has become a sharp tool for data mining,and has shown its capacity in many data mining competitions.By ensembling some base leaners with an effective strategy,ensemble learning usually can get a better model which has a good generalization in most cases.In this article,we have done the following works:We make some exploratory data analysis(EDA),data visualization and feature extraction on the dataset of taxi trip duration in New York City.This dataset is offered by NYC Taxi and Limousine Commission.We first train some rather good base leaners,and explore the ways to select them and combine them with ensemble methods to get a stronger model,so we can make a precise prediction of the trip duration.