Ensemble mobility predictor based on random forest and Markovian property using LBSN data

Abstract

The ubiquitous connectivity of Location-Based Systems (LBS) allows people to share individual location-related data anytime. In this sense, Location-Based Social Networks (LBSN) provides valuable information to be available in large-scale and low-cost fashion via traditional data collection methods. Moreover, this data contains spatial, temporal, and social features of user activity, enabling a system to predict user mobility. In this sense, mobility prediction plays crucial roles in urban planning, traffic forecasting, advertising, and recommendations, and has thus attracted lots of attention in the past decade. In this article, we introduce the Ensemble Random Forest-Markov (ERFM) mobility prediction model, a two-layer ensemble learner approach, in which the base learners are also ensemble learning models. In the inner layer, ERFM considers the Markovian property (memoryless) to build trajectories of different lengths, and the Random Forest algorithm to predict the user’s next location for each trajectory set. In the outer layer, the outputs from the first layer are aggregated based on the classification performance of each weak learner. The experimental results on the real user trajectory dataset highlight a higher accuracy and f1-score of ERFM compared to five state-of-the-art predictors.

Random forest using smartphone GPS in first wave of COVID-19 in the Maule region, Chile

“…Hence the importance of quarantines in the early stage of the pandemic; total mobility, active cases, external mobility, and internal mobility are the most relevant variables in VIM in the RF model used. Similar situations were observed in the studies in the Americas, Europe, and Africa (6)(7)(8). Larger and more populated cities have a lower external mobility index when the index is adjusted for population size.…”

Caching Policy in Low Earth Orbit Satellite Mega-Constellation Information-Centric Networking for Internet of Things

“…RF is one of the combined algorithms for data mining, reaching well-documented levels of accuracy and processing speed, and frequently appearing in new research [ 55 ]. RF as a supervised machine learning algorithm, uses a series of decision trees to distinguish between different classes of input data [ 35 , 56 ]. RF begins at the top of a decision tree and branches through a series of binary decisions, eventually leading to the input sample being defined as one of the possible categories [ 35 ], as shown in Figure 9 .…”

Predictive Fraud Detection: An Intelligent Method for Internet of Smart Grid Things Systems

“…These basic models often perform poorly by themselves because they have a high bias, such as low degree of freedom models, or because they have too much variance (e.g., high degree of freedom models). Hence, an ensemble predictor reduces the bias and/or variance of basic classifiers by combining several to create an aggregated learner (or ensemble predictor) that achieves better performances Araujo et al [2020].…”

“…In addition to being challenging to classify fraud, distinguishing and detecting different types of fraud is more complex using unique Machine Learning (ML) techniques. In this context, ensemble learning is an ML tech-nique where multiple predictors are trained to solve the same problem and combined to get better results Araujo et al [2020]. For instance, ML models often perform not so well by themselves either because they have a high bias or variance, and the idea of ensemble methods combining several of them to create an aggregated learner (or ensemble model) that achieves better performances.…”

“…Ensemble predictors, i.e., aggregated ML models, have become increasingly popular after showing excellent results in various research fields by combining the outputs of multiple ML algorithms to achieve better performance than a single classifier Araujo et al [2020]. The integration is more heterogeneous by considering the different time series classifiers types to classify these different parameters and attributes found in the user class.…”