Santiago Eibe

Clustering-based location in wireless networks

In this paper, we propose a three-phase methodology (measurement, calibration and estimation) for locating mobile stations (MS) in an indoor environment using wireless technology. Our solution is a fingerprint-based positioning system that overcomes the problem of the relative effect of doors and walls on signal strength and is independent of network device manufacturers. In the measurement phase, our system collects received signal strength indicator (RSSI) measurements from multiple access points. In the calibration phase, our system utilizes these measurements in a normalization process to create a radio map, a database of RSS patterns. Unlike traditional radio map-based methods, our methodology normalizes RSS measurements collected at different locations (on a floor) and uses artificial neural network models (ANNs) to group them into clusters. In the third phase, we use data mining techniques (clustering) to optimize location results. Experimental results demonstrate the accuracy of the proposed method. From these results it is clear that the system is highly likely to be able to locate a MS in a room or nearby room.

SOMAR: A SOcial Mobile Activity Recommender

A 2010 survey (Nielsen) showed that 22.7% of the time spent on the Internet is on a social network. Moreover, there is an increasing demand to access social networks by mobile phones, i.e., around 30% globally. Social networking has become a reality, and it generates an incredible amount of information that is sometimes difficult for users to process, especially from mobile phones. Several links, activities, and recommendations are proposed by networked friends every hour, which together are nearly impossible to manage. There is a need to filter and make accessible such information to users, which is the motivation behind developing a mobile recommender that exploits social network information. Thus, in this paper, we propose the design and the implementation of a SOcial Mobile Activity Recommender (SOMAR) that can integrate Facebook social network mobile data and sensor data to propose activities to the user (e.g., concert or computer science seminar). The recommendations are completely calculated in situ in the mobile device with an embedded data mining component. We analyze how to compute and update the social graph in case of changes in social relationships or user context. The paper also presents a case study to analyze the performance of the method.

Multi-agent location system in wireless networks

Highlights? A Multi-Agent Architecture and a methodology for indoor location is proposed. ? Mobile Station will have Fuzzy Location Agent (FLA) with minimum capacity processing. ? FLA establish its location on a plan of the floor of the building. ? FLA communicates with Fuzzy Location Manager Software Agent (FLMSA). ? FLMSA use fuzzy logic to estimation location based on a normalization process of RSS. In this paper we propose a flexible Multi-Agent Architecture together with a methodology for indoor location which allows us to locate any mobile station (MS) such as a Laptop, Smartphone, Tablet or a robotic system in an indoor environment using wireless technology. Our technology is complementary to the GPS location finder as it allows us to locate a mobile system in a specific room on a specific floor using the Wi-Fi networks.The idea is that any MS will have an agent known at a Fuzzy Location Software Agent (FLSA) with a minimum capacity processing at its disposal which collects the power received at different Access Points distributed around the floor and establish its location on a plan of the floor of the building. In order to do so it will have to communicate with the Fuzzy Location Manager Software Agent (FLMSA). The FLMSAs are local agents that form part of the management infrastructure of the Wi-Fi network of the Organization.The FLMSA implements a location estimation methodology divided into three phases (measurement, calibration and estimation) for locating mobile stations (MS). Our solution is a fingerprint-based positioning system that overcomes the problem of the relative effect of doors and walls on signal strength and is independent of the network device manufacturer.In the measurement phase, our system collects received signal strength indicator (RSSI) measurements from multiple access points. In the calibration phase, our system uses these measurements in a normalization process to create a radio map, a database of RSS patterns. Unlike traditional radio map-based methods, our methodology normalizes RSS measurements collected at different locations on a floor. In the third phase, we use Fuzzy Controllers to locate an MS on the plan of the floor of a building.Experimental results demonstrate the accuracy of the proposed method. From these results it is clear that the system is highly likely to be able to locate an MS in a room or adjacent room.

Self-configuring data mining for ubiquitous computing

Ubiquitous computing software needs to be autonomous so that essential decisions such as how to configure its particular execution are self-determined. Moreover, data mining serves an important role for ubiquitous computing by providing intelligence to several types of ubiquitous computing applications. Thus, automating ubiquitous data mining is also crucial. We focus on the problem of automatically configuring the execution of a ubiquitous data mining algorithm. In our solution, we generate configuration decisions in a resource aware and context aware manner since the algorithm executes in an environment in which the context often changes and computing resources are often severely limited. We propose to analyze the execution behavior of the data mining algorithm by mining its past executions. By doing so, we discover the effects of resource and context states as well as parameter settings on the data mining quality. We argue that a classification model is appropriate for predicting the behavior of an algorithms execution and we concentrate on decision tree classifier. We also define taxonomy on data mining quality so that tradeoff between prediction accuracy and classification specificity of each behavior model that classifies by a different abstraction of quality, is scored for model selection. Behavior model constituents and class label transformations are formally defined and experimental validation of the proposed approach is also performed.

A social network activity recommender system for ubiquitous devices

The increasing demand to access social networks by mobile devices together with increasing computation power of these mobile devices motivate the need of local recommendation services for social network users. Social networking is generating an incredible amount of information that is sometimes difficult for users to process, especially from mobile phones. Several links, activities, and recommendations are proposed by networked friends every hour, which together are nearly impossible to manage. There is a need to filter and make accessible such information to users, which is the motivation behind developing a mobile recommender that exploits social network information. Thus, in this paper, we propose the design and the implementation of a SOcial Mobile Activity Recommender (SOMAR) that can integrate Facebook social network mobile data and sensor data to propose activities to the user. The recommendations are completely calculated in situ in the mobile device with an embedded data mining component that is the basis to compute a social graph of the user relationships that will be later used for the recommendation process. The paper also presents some experiments that analyze the performance of the proposed method.

Dynamic Clustering Process to Calculate Affinity Degree of Users as Basis of a Social Network Recommender

Social networking has become a reality: links, activities, and recommendations are proposed by networked friends every moment. There is a need to filter such information to make user enjoy such an experience. In the same way recommender systems where proposed to ease the browsing experience of navigators, nowadays recommenders are required to help users in sharing and obtaining the appropriate information on the social networks. The challenges behind are not only related to the continuous evolution of information being shared, but also by the fact that ubiquity is today a reality. Consequently recommender should take into account the context of the user to whom recommendations are being done. In this paper we present a recommender for social networks that is able to update recommendations as information evolves. The recommender is based on a graph build on basis of a data mining component that extract knowledge on relations and information exchanged by users. The mining component can run autonomously so recommendations can be updated if required. The paper also presents preliminary analysis on the performance of the proposed recommender.

Adapting Batch Learning Algorithms Execution in Ubiquitous Devices

In order to provide context aware, adaptive, and anticipatory services, data mining services are required to provide them with intelligence. The data mining could be either executed in a central server or locally. In either case, adaptability to the changing environment is required. In the stream mining scenario, some solutions have been proposed to provide mechanisms to adapt the execution to available resources and context. Here, we propose a cost model mechanism to adapt the algorithm execution according to available resources and context information for the case of static data. The mechanism based on analyzing efficacy and efficiency (EE-Model) of the algorithm, is a two step process in which first the efficiency and efficacy of the algorithm are calculated for predefined algorithm configurations and dataset input. In a second step, taking into account the available resources and context, the best configuration of the algorithm is chosen. The paper describes the mechanism and presents an EE-Model instantiation for C4.5 algorithm. Further, we demonstrate the convenience of the proposed approach with a simulation of synthetic data.

Situation-Aware Data Mining Service for Ubiquitous Environments

The indisputable dominance of mobile and pervasive computing devices and their typical characteristics require services offered to be rethought and sometimes redesigned in order to better assist users. Considering the importance of data mining services to provide intelligence locally on devices on these environments, we propose a data mining service that adapts the embedded data mining algorithm according to situation. Resource-awareness and context-awareness are the essential features that the proposed service will have to provide. Consequently we present a model in which data mining configuration is determined based on context and resources. We separate control and functionality in order to provide more flexibility and comply with existing data mining standards. An adaptable design is attained through definition of situations and strategies. The mechanism used in definition of strategies is an important factor affecting the performance of the control part which determines the configuration of data mining algorithm. Anticipating the importance of the mechanism selection, the paper also presents comparison with three different mechanisms. We designed a situation-aware data mining service favoring adaptability and efficiency as the important features and assessed the alternative representations of its components.

Bayesian networks to predict data mining algorithm behavior in ubiquitous computing environments

The growing demand of data mining services for ubiquitous computing environments necessitates deployment of appropriate mechanisms that make use of circumstantial factors to adapt the data mining behavior. Despite the efforts and results so far for efficient parameter tuning, incorporating dynamically changing context information on the parameter setting decision is lacking in the present work. Thus, Bayesian networks are used to learn, in possible situations the effects of data mining algorithm parameters on the final model obtained. Based on this knowledge, we propose to infer future algorithm configurations appropriate for situations. Instantiation of the approach for association rules is also shown in the paper and the feasibility of the approach is validated by the experimentation.

Autonomous adaptive data mining for u-healthcare

Ubiquitous healthcare requires intelligence in order to be able to react to different patients needs. The context and resources constraints of the ubiquitous devices demand a mechanism able to estimate the cost of the data mining algorithm providing the intelligence. The performance of the algorithm is independent of the semantics, this is to say, knowing the input of an algorithm the performance can be calculated. Under this assumption we present formalization of a mechanism able to estimate the cost of an algorithm in terms of efficacy and efficiency. Further, an instantiation of the mechanism for an application predicting glucose level for diabetic patients is presented.