| Pub. Date | : Jan, 2020 |
|---|---|
| Product Name | : The IUP Journal of Applied Economics |
| Product Type | : Article |
| Product Code | : IJAE40120 |
| Author Name | : Sanjay Fuloria |
| Availability | : YES |
| Subject/Domain | : Economics |
| Download Format | : PDF Format |
| No. of Pages | : 15 |
Forecasting of passenger demand in the short term could be beneficial to the ondemand ride services platforms. This would help the platforms to incentivize drivers by moving them from areas of low demand to high demand. This, in turn, would help the platform companies to maximize their profits too. Using Uber dataset for New York City, the present study compares three different methods-exponential smoothing, multiple regression and Long Short-Term Memory (LSTM)-to forecast passenger demand in the short term. While traditional statistical methods like exponential smoothing and multiple regression are more explainable, deep learning methods like LSTM, with their complex methodologies, are more accurate in some situations. Additionally, temperature dataset for the city of New York is also used for forecasting. The study concludes that LSTM models perform better in forecasting both for training and validation datasets. The research could be further enhanced by using a bigger and geographically diverse dataset.
Ridesharing services like Uber, Ola, and Lyft use technology to create a platform where cab drivers and customers meet for mutual benefit. In this setup, private car owners provide rides to customers for a fee (Chen et al., 2017). Ridesharing services are accessed through a smartphone application. The platform plays the matchmaker by matching supply and demand. The registered drivers constitute the supply side and the customers looking for rides constitute the demand side in this case. There are a lot of factors that influence both the drivers' and customers' behaviors and preferences. These behaviors and preferences impact the demand and supply. This in turn affects the profitability of the platforms and the economic wellbeing of the drivers. In such a situation, it is imperative to understand the short-term passenger demand across a city for an operator. This would help the operator to incentivize the cab drivers by reallocating them from an area of low demand to high demand.