João Bártolo Gomes

MARS: A Personalised Mobile Activity Recognition System

Mobile activity recognition focuses on inferring the current activities of a mobile user by leveraging the sensory data that is available on todays smart phones. The state of the art in mobile activity recognition uses traditional classification techniques. Thus, the learning process typically involves: i) collection of labelled sensory data that is transferred and collated in a centralised repository, ii) model building where the classification model is trained and tested using the collected data, iii) a model deployment stage where the learnt model is deployed on-board a mobile device for identifying activities based on new sensory data. In this paper, we demonstrate the Mobile Activity Recognition System (MARS) where for the first time the model is built and continuously updated on-board the mobile device itself using data stream mining. The advantages of the on-board approach are that it allows model personalisation and increased privacy as the data is not sent to any external site. Furthermore, when the user or its activity profile changes MARS enables quick model adaptation. One of the stand out features of MARS is that training/updating the model takes less than 30 seconds per activity. MARS has been implemented on the Android platform to demonstrate that it can achieve accurate mobile activity recognition. Moreover, we can show in practice that MARS quickly adapts to user profile changes while at the same time being scalable and efficient in terms of consumption of the device resources.

Learning recurring concepts from data streams with a context-aware ensemble

The dynamic and unstable nature observed in real world applications influences learning systems through changes in data, context and resource availability. Data stream mining systems must be aware and adapt to such changes so that incoming data can continuously be classified with high accuracy. Ensemble approaches have been shown successful in dealing with concept changes. Despite their success in learning under concept changes, context information has not yet been exploited by ensemble approaches in data stream scenarios where concepts reappear. Under these circumstances, context information appropriately integrated with learned concepts would enable to anticipate recurring changes in concepts. In this work, we present an ensemble based approach for the problem of detecting concept changes in data streams where concepts reappear, that dynamically adds and removes weighted classifiers in response to changes not only in concepts but to context. We identify stable concepts using a change detection method, based on the error-rate of the learning process. Context information is used in the adaptation to recurring concepts and in the management of knowledge from previous learned concepts while adapting to resource constraints. Consequently, proper representation and storage of context and concepts is a major issue dealt within the paper. We present and discuss preliminary experimental results with synthetic and real datasets.

An overview of state-of-the-art partial discharge analysis techniques for condition monitoring

As one step toward the future smart grid, condition monitoring is important to facilitate the reliability of grid asset operation and to save on maintenance cost [1]. Most failures of the power grid are caused by electrical insulation failure, and a key indicator of such electrical failure is the occurrence of partial discharge (PD). Therefore, one focus of condition monitoring is to detect PD, especially in the early stages, to prevent a serious power failure or outage.

Mining Recurring Concepts in a Dynamic Feature Space

Most data stream classification techniques assume that the underlying feature space is static. However, in real-world applications the set of features and their relevance to the target concept may change over time. In addition, when the underlying concepts reappear, reusing previously learnt models can enhance the learning process in terms of accuracy and processing time at the expense of manageable memory consumption. In this paper, we propose mining recurring concepts in a dynamic feature space (MReC-DFS), a data stream classification system to address the challenges of learning recurring concepts in a dynamic feature space while simultaneously reducing the memory cost associated with storing past models. MReC-DFS is able to detect and adapt to concept changes using the performance of the learning process and contextual information. To handle recurring concepts, stored models are combined in a dynamically weighted ensemble. Incremental feature selection is performed to reduce the combined feature space. This contribution allows MReC-DFS to store only the features most relevant to the learnt concepts, which in turn increases the memory efficiency of the technique. In addition, an incremental feature selection method is proposed that dynamically determines the threshold between relevant and irrelevant features. Experimental results demonstrating the high accuracy of MReC-DFS compared with state-of-the-art techniques on a variety of real datasets are presented. The results also show the superior memory efficiency of MReC-DFS.

Data stream mining in ubiquitous environments: state-of-the-art and current directions

In this article, we review the state‐of‐the‐art techniques in mining data streams for mobile and ubiquitous environments. We start the review with a concise background of data stream processing, presenting the building blocks for mining data streams. In a wide range of applications, data streams are required to be processed on small ubiquitous devices like smartphones and sensor devices. Mobile and ubiquitous data mining target these applications with tailored techniques and approaches addressing scarcity of resources and mobility issues. Two categories can be identified for mobile and ubiquitous mining of streaming data: single‐node and distributed. This survey will cover both categories. Mining mobile and ubiquitous data require algorithms with the ability to monitor and adapt the working conditions to the available computational resources. We identify the key characteristics of these algorithms and present illustrative applications. Distributed data stream mining in the mobile environment is then discussed, presenting the Pocket Data Mining framework. Mobility of users stimulates the adoption of context‐awareness in this area of research. Context‐awareness and collaboration are discussed in the Collaborative Data Stream Mining, where agents share knowledge to learn adaptive accurate models.

CARDAP: A scalable energy-efficient context aware distributed mobile data analytics platform for the fog

Distributed online data analytics has attracted significant research interest in recent years with the advent of Fog and Cloud computing. The popularity of novel distributed applications such as crowdsourcing and crowdsensing have fostered the need for scalable energy-efficient platforms that can enable distributed data analytics. In this paper, we propose CARDAP, a (C)ontext (A)ware (R)eal-time (D)ata (A)nalytics (P)latform. CARDAP is a generic, flexible and extensible, component- based platform that can be deployed in complex distributed mobile analytics applications e.g. sensing activity of citizens in smart cities. CARDAP incorporates a number of energy efficient data delivery strategies using real-time mobile data stream mining for data reduction and thus less data transmission. Extensive experimental evaluations indicate the CARDAP platform can deliver significant benefits in energy efficiency over naive approaches. Lessons learnt and future work conclude the paper.

Where Will You Go? Mobile Data Mining for Next Place Prediction

The technological advances in smartphones and their widespread use has resulted in the big volume and varied types of mobile data which we have today. Location prediction through mobile data mining leverages such big data in applications such as traffic planning, location-based advertising, intelligent resource allocation; as well as in recommender services including the popular Apple Siri or Google Now. This paper, focuses on the challenging problem of predicting the next location of a mobile user given data on his or her current location. In this work, we propose NextLocation - a personalised mobile data mining framework - that not only uses spatial and temporal data but also other contextual data such as accelerometer , bluetooth and call/sms log. In addition, the proposed framework represents a new paradigm for privacy-preserving next place prediction as the mobile phone data is not shared without user permission. Experiments have been performed using data from the Nokia Mobile Data Challenge MDC. The results on MDC data show large variability in predictive accuracy of about 17% across users. For example, irregular users are very difficult to predict while for more regular users it is possible to achieve more than 80% accuracy. To the best of our knowledge, our approach achieves the highest predictive accuracy when compared with existing results.

A rule dynamics approach to event detection in Twitter with its application to sports and politics

We applied Association Rule Mining on tweets hashtags.We used hashtag keywords for Topic Detection & Tracking on Twitter.We performed our experiments on datasets from sports and political domains. The increasing popularity of Twitter as social network tool for opinion expression as well as information retrieval has resulted in the need to derive computational means to detect and track relevant topics/events in the network. The application of topic detection and tracking methods to tweets enable users to extract newsworthy content from the vast and somehow chaotic Twitter stream. In this paper, we apply our technique named Transaction-based Rule Change Mining to extract newsworthy hashtag keywords present in tweets from two different domains namely; sports (The English FA Cup 2012) and politics (US Presidential Elections 2012 and Super Tuesday 2012). Noting the peculiar nature of event dynamics in these two domains, we apply different time-windows and update rates to each of the datasets in order to study their impact on performance. The performance effectiveness results reveal that our approach is able to accurately detect and track newsworthy content. In addition, the results show that the adaptation of the time-window exhibits better performance especially on the sports dataset, which can be attributed to the usually shorter duration of football events.

Scalable Energy-Efficient Distributed Data Analytics for Crowdsensing Applications in Mobile Environments

We are witnessing a new revolution in computing and communication involving symbiotic networks of people (social networks), intelligent devices, smart mobile computing, and communication devices that will form cyber-physical social systems. The emergence of intelligent devices with monitoring, sensing, and actuation capabilities referred to as Internet of Things and social networks have increased the popularity of novel social applications such as crowdsourcing and crowdsensing. The upsurge of such applications has fostered the need for scalable cost-efficient platforms that can enable distributed data analytics. In this paper, we propose CARDAP, a scalable, energy-efficient, generic and extensible component-based distributed data analytics platform for mobile crowdsensing (MCS) applications. CARDAP incorporates on-the-move activity recognition and a number of energy efficient data delivery strategies using real-time mobile data stream mining. We propose and develop theoretical cost models for typical crowdsensing application scenarios. Experimental evaluations of CARDAP using a proof-of-concept MCS scenario validate the theoretical cost model estimates and demonstrate the platforms ability to deliver significant benefits in energy, resource, and query processing efficiency.

CALDS: context-aware learning from data streams

Drift detection methods in data streams can detect changes in incoming data so that learned models can be used to represent the underlying population. In many real-world scenarios context information is available and could be exploited to improve existing approaches, by detecting or even anticipating to recurring concepts in the underlying population. Several applications, among them health-care or recommender systems, lend themselves to use such information as data from sensors is available but is not being used. Nevertheless, new challenges arise when integrating context with drift detection methods. Modeling and comparing context information, representing the context-concepts history and storing previously learned concepts for reuse are some of the critical problems. In this work, we propose the Context-aware Learning from Data Streams (CALDS) system to improve existing drift detection methods by exploiting available context information. Our enhancement is seamless: we use the association between context information and learned concepts to improve detection and adaptation to drift when concepts reappear. We present and discuss our preliminary experimental results with synthetic and real datasets.