scholarly journals Location Prediction with Personalized Federated Learning

Author(s):  
shuang wang ◽  
Bowei Wang ◽  
Shuai Yao ◽  
Jiangqin Qu ◽  
Yuezheng Pan

Abstract Location prediction has attracted wide attention in human mobility prediction because of the popularity of location-based social networks. Existing location prediction methods have achieved remarkable development in centrally stored datasets. However, these datasets contain privacy data about user behaviors and may cause privacy issues. A location prediction method is proposed in our work to predict human movement behavior using federated learning techniques in which the data is stored in different clients and different clients cooperate to train to extract useful users’ behavior information and prevent the disclosure of privacy. Firstly, we put forward an innovative spatial-temporal location prediction framework(STLPF) for location prediction by integrating spatial-temporal information in local and global views on each client, and propose a new loss function to optimize the model. Secondly, we design a new personalized federated learning framework in which clients can cooperatively train their personalized models in the absence of a global model. Finally, the numerous experimental results on check-in datasets further show that our privacy-protected method is superior and more effective than various baseline approaches.

Author(s):  
R. Javanmard ◽  
R. Esmaeili ◽  
M. Malekzadeh ◽  
F. Karimipour

Abstract. Movement data are becoming extensive and comprehensive with the advent of GPS (global positioning system) and pervasive use of smartphones, which has led to an increasing rate of studies about movement such as mobility pattern of oil spills, taxies, storms and animals. Studying the movement of people has long been the topic of much thought and debate among researchers within the field of transportation, social issues, and policy. One of the basic prerequisites for studying human movement behavior is modeling the movement, which show how people move so that the effect of different variables can be revealed. For this purpose, this research intends to deploy the concept of activity space (i.e., the part of the space in which a person is active) and its determinants to display the trajectory of individuals, and then modeling the effect of different variables on human mobility behavior. This study explores the effect of time (movement on weekends and weekdays) and demographic (age, gender, occupation state) factors on the characteristics of human mobility pattern and analyzes the extent to which the mobility pattern of different group of people is related to time by using Swiss human movement sample dataset, called MDC. These movement characteristics can be used later in a wide range of applications, such as predictions, urban planning, and traffic forecasting.


2020 ◽  
Author(s):  
Mohammad Alarifi ◽  
Somaieh Goudarzvand3 ◽  
Abdulrahman Jabour ◽  
Doreen Foy ◽  
Maryam Zolnoori

BACKGROUND The rate of antidepressant prescriptions is globally increasing. A large portion of patients stop their medications which could lead to many side effects including relapse, and anxiety. OBJECTIVE The aim of this was to develop a drug-continuity prediction model and identify the factors associated with drug-continuity using online patient forums. METHODS We retrieved 982 antidepressant drug reviews from the online patient’s forum AskaPatient.com. We followed the Analytical Framework Method to extract structured data from unstructured data. Using the structured data, we examined the factors associated with antidepressant discontinuity and developed a predictive model using multiple machine learning techniques. RESULTS We tested multiple machine learning techniques which resulted in different performances ranging from accuracy of 65% to 82%. We found that Radom Forest algorithm provides the highest prediction method with 82% Accuracy, 78% Precision, 88.03% Recall, and 84.2% F1-Score. The factors associated with drug discontinuity the most were; withdrawal symptoms, effectiveness-ineffectiveness, perceived-distress-adverse drug reaction, rating, and perceived-distress related to withdrawal symptoms. CONCLUSIONS Although the nature of data available at online forums differ from data collected through surveys, we found that online patients forum can be a valuable source of data for drug-continuity prediction and understanding patients experience. The factors identified through our techniques were consistent with the findings of prior studies that used surveys.


2021 ◽  
Vol 21 (1) ◽  
Author(s):  
Meng-Chun Chang ◽  
Rebecca Kahn ◽  
Yu-An Li ◽  
Cheng-Sheng Lee ◽  
Caroline O. Buckee ◽  
...  

Abstract Background As COVID-19 continues to spread around the world, understanding how patterns of human mobility and connectivity affect outbreak dynamics, especially before outbreaks establish locally, is critical for informing response efforts. In Taiwan, most cases to date were imported or linked to imported cases. Methods In collaboration with Facebook Data for Good, we characterized changes in movement patterns in Taiwan since February 2020, and built metapopulation models that incorporate human movement data to identify the high risk areas of disease spread and assess the potential effects of local travel restrictions in Taiwan. Results We found that mobility changed with the number of local cases in Taiwan in the past few months. For each city, we identified the most highly connected areas that may serve as sources of importation during an outbreak. We showed that the risk of an outbreak in Taiwan is enhanced if initial infections occur around holidays. Intracity travel reductions have a higher impact on the risk of an outbreak than intercity travel reductions, while intercity travel reductions can narrow the scope of the outbreak and help target resources. The timing, duration, and level of travel reduction together determine the impact of travel reductions on the number of infections, and multiple combinations of these can result in similar impact. Conclusions To prepare for the potential spread within Taiwan, we utilized Facebook’s aggregated and anonymized movement and colocation data to identify cities with higher risk of infection and regional importation. We developed an interactive application that allows users to vary inputs and assumptions and shows the spatial spread of the disease and the impact of intercity and intracity travel reduction under different initial conditions. Our results can be used readily if local transmission occurs in Taiwan after relaxation of border control, providing important insights into future disease surveillance and policies for travel restrictions.


2021 ◽  
Vol 7 (4) ◽  
pp. 1-24
Author(s):  
Douglas Do Couto Teixeira ◽  
Aline Carneiro Viana ◽  
Jussara M. Almeida ◽  
Mrio S. Alvim

Predicting mobility-related behavior is an important yet challenging task. On the one hand, factors such as one’s routine or preferences for a few favorite locations may help in predicting their mobility. On the other hand, several contextual factors, such as variations in individual preferences, weather, traffic, or even a person’s social contacts, can affect mobility patterns and make its modeling significantly more challenging. A fundamental approach to study mobility-related behavior is to assess how predictable such behavior is, deriving theoretical limits on the accuracy that a prediction model can achieve given a specific dataset. This approach focuses on the inherent nature and fundamental patterns of human behavior captured in that dataset, filtering out factors that depend on the specificities of the prediction method adopted. However, the current state-of-the-art method to estimate predictability in human mobility suffers from two major limitations: low interpretability and hardness to incorporate external factors that are known to help mobility prediction (i.e., contextual information). In this article, we revisit this state-of-the-art method, aiming at tackling these limitations. Specifically, we conduct a thorough analysis of how this widely used method works by looking into two different metrics that are easier to understand and, at the same time, capture reasonably well the effects of the original technique. We evaluate these metrics in the context of two different mobility prediction tasks, notably, next cell and next distinct cell prediction, which have different degrees of difficulty. Additionally, we propose alternative strategies to incorporate different types of contextual information into the existing technique. Our evaluation of these strategies offer quantitative measures of the impact of adding context to the predictability estimate, revealing the challenges associated with doing so in practical scenarios.


2021 ◽  
Vol 12 (6) ◽  
pp. 1-23
Author(s):  
Shuo Tao ◽  
Jingang Jiang ◽  
Defu Lian ◽  
Kai Zheng ◽  
Enhong Chen

Mobility prediction plays an important role in a wide range of location-based applications and services. However, there are three problems in the existing literature: (1) explicit high-order interactions of spatio-temporal features are not systemically modeled; (2) most existing algorithms place attention mechanisms on top of recurrent network, so they can not allow for full parallelism and are inferior to self-attention for capturing long-range dependence; (3) most literature does not make good use of long-term historical information and do not effectively model the long-term periodicity of users. To this end, we propose MoveNet and RLMoveNet. MoveNet is a self-attention-based sequential model, predicting each user’s next destination based on her most recent visits and historical trajectory. MoveNet first introduces a cross-based learning framework for modeling feature interactions. With self-attention on both the most recent visits and historical trajectory, MoveNet can use an attention mechanism to capture the user’s long-term regularity in a more efficient way. Based on MoveNet, to model long-term periodicity more effectively, we add the reinforcement learning layer and named RLMoveNet. RLMoveNet regards the human mobility prediction as a reinforcement learning problem, using the reinforcement learning layer as the regularization part to drive the model to pay attention to the behavior with periodic actions, which can help us make the algorithm more effective. We evaluate both of them with three real-world mobility datasets. MoveNet outperforms the state-of-the-art mobility predictor by around 10% in terms of accuracy, and simultaneously achieves faster convergence and over 4x training speedup. Moreover, RLMoveNet achieves higher prediction accuracy than MoveNet, which proves that modeling periodicity explicitly from the perspective of reinforcement learning is more effective.


Author(s):  
Nina Glick Schiller

Debates about migration, whether led by politicians or scholars, often approach migration as a relatively new challenge and categorize it as a “destabilizing force,” ignoring the fact that the world’s past and present has been built by human movement. Humans have always migrated. Individual and population mobility as well as settlement are part of humans’ shared history. To integrate migration into an understanding of humans’ shared past, present, and emerging possible futures, several concepts prove useful including migration regime, displacement, dispossession, conjuncture, colonization, border-making, nationalism, and racialization. Deployed together, these concepts identify moments in human history in which migration has been understood to be part of the human experience and when, where, and how migrants have been stigmatized, and those who move defined as culturally or biologically inferior. By coupling the concept of migration regimes with an analysis of changing modes of dispossession and displacement over millennia, scholars can illuminate the intersection of the economic and political transformations of governance structures as well as the varying concepts of “the migrant” and “nonmigrant,” and “native” and “foreigner.” Anti-immigrant ideologies preclude discussion of the broader economic and political restructurings that underlie both increased human movement and anti-migrant sentiments. They also deflect attention from a set of questions that are at the heart of the anthropology of migration: Why do people leave familiar terrains, family, and friends? How do they manage to move and settle elsewhere? How do they relate to the life they left behind? These are questions that can equally be asked of people who move to another region of a country or travel across political boundaries. To answer these questions migration scholars have explored the linkages between forms of human mobility and processes of dispossession, displacement, and resettlement. In these investigations, social networks prove to be central to mobility and settlement. Since the 15th century, changing Western theories about human migration and the origins of political and social boundaries reflected transformations in political economy. Globe-spanning migration regimes used violent force, border formation and dissolution, documents, surveillance, and criminalization to allow the migration of some and disallow the movement or settlement of others. During that period, marked initially by colonialism and slavery, and then by nation state building and anticolonial struggles, migration scholars including the anthropologists took varying positions on the significance of mobility and stasis in human life. By the beginning of the 21st century, the accumulation of capital by dispossession emerged as a process increasingly central to a historical conjuncture marked by both heightened migration and anti-immigrant nationalism. Political struggles for social and environmental justice began to merge with movements in support of migration. This political climate shaped a new engaged anthropology of migration.


PLoS ONE ◽  
2018 ◽  
Vol 13 (11) ◽  
pp. e0207063 ◽  
Author(s):  
Yongping Du ◽  
Chencheng Wang ◽  
Yanlei Qiao ◽  
Dongyue Zhao ◽  
Wenyang Guo

Sign in / Sign up

Export Citation Format

Share Document