Predicting Taxi–Passenger Demand Using Streaming Data
Citations
507 citations
Cites background from "Predicting Taxi–Passenger Demand Us..."
...Moreira-Matias et al. (2013) proposed a data stream ensemble framework which incorporated time varying passion model and ARIMA, to predict the spatial distribution of taxi passenger demand....
[...]
...However, heterogeneous and exogenous factors in reality, e.g., asymmetric information, short-term fluctuations, may make it difficult to guarantee the spatial distribution of taxis matching the passenger demand all the time (Moreira-Matias et al., 2013)....
[...]
..., asymmetric information, short-term fluctuations, may make it difficult to guarantee the spatial distribution of taxis matching the passenger demand all the time (Moreira-Matias et al., 2013)....
[...]
408 citations
311 citations
303 citations
302 citations
References
272,030 citations
"Predicting Taxi–Passenger Demand Us..." refers background in this paper
...Finally, the results that are achieved are presented and discussed....
[...]
[...]
9,919 citations
"Predicting Taxi–Passenger Demand Us..." refers background or methods in this paper
...2....
[...]
...To do so, well-known time-series forecasting techniques were used and adapted to this problem, such as the time-varying Poisson model [15] and the autoregressive integrated moving average (ARIMA) [16]....
[...]
...Third, they presented an improved ARIMA depending both on time and day type....
[...]
...The ARIMA model (p, d, and q values, and seasonality) was first set (and updated each 24 h) by learning/detecting the underlying model (i.e., autocorrelation and partial autocorrelation analysis) running on the historical time-series curve of each stand during the last two weeks (i.e., period t− 2θ, t)....
[...]
...Despite their good results, this approach comparatively has the following three weak points to the one presented: 1) it just uses the most immediate historical data, discarding the mid- and long-term memory of the system; 2) in their testbed, the authors use minimum aggregation periods of 60 min over offline historical data (i.e., the next value prediction task on a time series is easier as long as the aggregation period is increased), whereas we use short-term periods of 30 min; and 3) the work does not clearly describe how the authors update both the ARIMA model and the weights that are used by it....
[...]
5,514 citations
4,641 citations
3,155 citations
Additional excerpts
...A brief presentation of one of the simplest ARIMA models (for nonseasonal stationary time series) is presented next, following the existing description in [30] (however, our framework can...
[...]