International Research Journal of Engineering and Technology (IRJET)
e-ISSN: 2395-0056
Volume: 10 Issue: 05 | May 2023
p-ISSN: 2395-0072
www.irjet.net
Taxi Demand Prediction using Machine Learning. P. Sudheer Benarji1, P. Sai Bharadwaj2, B. Neeha3, D. Srikanth4, V. Ankitha5 1Professor, VNR Vignana Jyothi Institute of Engineering & Technology, Hyderabad
2345Under Graduate Student, VNR Vignana Jyothi Institute of Engineering & Technology, Hyderabad
---------------------------------------------------------------------***--------------------------------------------------------------------prediction problem, this study develops a Graph MultiAbstract - Taxi demand prediction is the process of using historical data to forecast future taxi requests in a particular area. Managers may pre-allocate taxi resources in cities with the aid of accurate and real-time demand forecasting, which helps drivers find clients more quickly and cuts down on passenger waiting times. This project is aimed to choose the best model in predicting the taxi demand where we use various Machine learning techniques such as regression analysis and time series forecasting. Various baseline models, including moving averages (simple, weighted, and exponential), linear regression with grid search, random forest regressor with random search, and XGBoost regressor with random search are used. We find out which model is more suitable in predicting the output using the metrics we obtain.
Key Words: Linear Regression with GridSearchCV, Random Forest Regressor with RandomSearchCV, XGBoost Regressor with RandomSearchCV.
1.INTRODUCTION Taxi demand prediction is the process of using historical data to forecast future taxi requests in a particular area. Managers may pre-allocate taxi resources in cities with the aid of accurate and real-time demand forecasting, which helps drivers find clients more quickly and cuts down on passenger waiting times. In our project, we’ve used data on taxi rides in New York city to to train and test the models using Linear regression, Random Forest regressor and XGBoost regressor. Along with the Linear regression and Random Forest algorithms, we’ve also used the XGBoost algorithm. XGBoost is a machine learning algorithm that is commonly used in classification and regression problems. It is an ensemble learning method that combines the weak prediction models , such as decision trees to create a stronger overall prediction model. XGBoost has gained popularity due to its high accuracy, scalability, and ability to handle missing data. With our project, we get an understanding of which model is best to predict the real time taxi demand and taxi companies be able to tailor strategies to allocate resources based on demand.
2. Literature Review Multi-attention network-based graph prediction of taxi demand: In order to better address the taxi demand
© 2023, IRJET
|
Impact Factor value: 8.226 |
Attention Network (GMAN), which tries to forecast the taxi demands in every section of a road network.(Achieved a 72% Accuracy). Because only significant data needs to be learned by the models, applying attention increases model accuracy to extremely high levels. The Attention mechanism's drawback is that it requires a lot of time and is challenging to parallelize Taxi demand forecast using the random forest model: Decision trees are employed in the random forest. The term "random" refers to our usage of a random bootstrap sampling, and the term "forest" refers to the collection of trees seen in decision trees. (Achieved 77% Accuracy). excellent forecasting abilities that improve application precision. Easy data preparation is made possible by not requiring normalization. Generally speaking, this algorithm is quick to train but takes a while to produce predictions after training
Demand projection for taxis XGBoost algorithmbased: The hot spot locations are identified and their boundaries are drawn using the density-based DBSCAN clustering technique, and the demand for the hot spot areas is predicted using the XGBoost algorithm. XGB provides various features, such as parallelization, cache optimization, and more. Like any other boosting method, XGBoost is sensitive to outliers. Taxi Demand Forecast Based on Regional Heterogeneity Analysis and Multi-Level Deep Learning: With the aid of the taxi zone clustering technique and pairwise clustering theory, the Multi-Level Recurrent Neural Networks (MLRNN) model is put out.(83.33% Accuracy Attained). concentrates on how to exploit inter-zone heterogeneity to enhance prediction. The use of MLRNN results in high processing costs and greater complexity when fitting data. Probabilistic Taxi Demand Prediction with Bayesian Deep Learning: Proposes a Bayesian deep learning approach for probabilistic taxi demand prediction. (Achieved Accuracy of 83%). Estimates the uncertainty of predictions and provides probabilistic forecasts. Greater technical complexity, defining a prior distribution can be hard using Bayesian statistics. Prediction of Taxi Demand Using Ensemble Model: Utilizes a point of interest (POI) to match taxi demand with a location so that it can be studied using a different function.
ISO 9001:2008 Certified Journal
|
Page 1527