MMARRS: An Intelligent Route Recommendation and Road Traffic Information System for Multimodal and Unimodal Public Transportation using Text Analysis

Public commuting in the Philippines, particularly in the Metro Manila setting, continuously rises as a crucial problem due to also constantly voluminously increasing traffic congestion. Hence, this study proposes an intelligent route recommendation and road traffic information system that uses LTFRB data on 1) fare matrix computation (LRT 1, LRT 2, MRT 3, Bus, and Jeep) through utilizing Sakay.ph GTFS data, as well as 2) taxi pricing computation, and MMDA data (from their official Twitter account) on real-time traffic situation. Such data are used 1) to process the top three route recommendation choices, 2) to create attribute-based bag of words, extract appropriate dataset features, and classify the traffic congestion mode (Light, Light to Moderate, Moderate, Moderate to Heavy, and Heavy) in the involved road/s using Latent Dirichlet Allocation (LDA), and 3) to rebuild the system model automatically in a certain time interval. In this study, 1) the traffic-related tweets from the official Twitter account of MMDA are fetched using Twitter Streaming API and filtered using Named Entity Recognition; 2) the filtered data are preprocessed by applying tokenization, frequency counting, and removal of unnecessary symbols; 3) the features from the preprocessed data are then extracted using Latent Dirichlet Allocation and are hereby used to identify the significant topic segments (time, day, lane of road, road direction, location and traffic mode); and 4) Linear Regression was used for pattern recognition. The results found are as follows: 1) 84% for the accuracy, 85% for the precision, and 83% for the recall garnered for the applied methodology using k-NN as the chosen classification model; 2) the advantage of supervised data acquisition over unsupervised data acquisition; and 3) traffic mode-based pattern extraction and evaluation. These results show the usability and practicality of the study to public commuting.