User context recognition using smartphone sensors and classification models

Otebolaku, Abayomi Moradeyo; Andrade, Maria Teresa

doi:10.1016/j.jnca.2016.03.013

Cited by 46 publications

(71 citation statements)

References 35 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In their raw forms, they are not suitable for building real-world applications. Noise, drifts, and delays are some of the common sources of sensor data errors [12] [14]. In order to mitigate the influence of these errors and noise in the raw data, since they can corrupt the captured context information and consequently the inferred contextual situation, raw data filtering must be performed.…”

Section: (I) Context Sensingmentioning

confidence: 99%

“…Let r be the length of the windows slide. For a 50% windows slide, r = 0.5 l. Using these definitions, function f1 is defined as follows in equation (3) The second phase of the feature extraction process involves generating a set of statistical features from each window known as feature extraction [5] [12]. Following the sliding windowing process explained in the previous section, statistical features are generated from each of the matrices Mxi, Myi, Mzi to build vector Vi with labeled contexts a∈A = {a1, a2,…,an}, of the agent.…”

Section: A) Sliding Window Phasementioning

confidence: 99%

See 1 more Smart Citation

Towards context classification and reasoning in IoT

Otebolaku

Lee

2017

2017 14th International Conference on Telecommunications (ConTEL)

Self Cite

View full text Add to dashboard Cite

Abstract-Internet of Things (IoT) is the future of ubiquitous and personalized intelligent service delivery. It consists of interconnected, addressable and communicating everyday objects. To realize the full potentials of this new generation of ubiquitous systems, IoT's 'smart' objects should be supported with intelligent platforms for data acquisition, pre-processing, classification, modeling, reasoning and inference including distribution. However, some current IoT systems lack these capabilities: they provide mainly the functionality for raw sensor data acquisition. In this paper, we propose a framework towards deriving high-level context information from streams of raw IoT sensor data, using artificial neural network (ANN) as context recognition model. Before building the model, raw sensor data were pre-processed using weighted average low-pass filtering and a sliding window algorithm. From the resulting windows, statistical features were extracted to train ANN models. Analysis and evaluation of the proposed system show that it achieved between 87.3% and 98.1% accuracies.

show abstract

Section: (I) Context Sensingmentioning

confidence: 99%

Section: A) Sliding Window Phasementioning

confidence: 99%

Towards context classification and reasoning in IoT

Otebolaku

Lee

2017

2017 14th International Conference on Telecommunications (ConTEL)

Self Cite

View full text Add to dashboard Cite

show abstract

“…This is possible because we have developed a dynamic context recognition model as an integral part of our broader personalization system [24,28,31]. First, the model acquires low-level context data from the user's handheld device to classify and identify their contextual situations.…”

Section: A High-level View Of New User's Preference Prediction Using mentioning

confidence: 99%

“…The detailed description of how this was realized is out of the scope of the present article, interested readers can take a look at Ref. [24] for details. Second, having identified the current context of the new user, the system then uses this context to search the profiles of existing users to identify those users with contexts that are similar to the new user's context.…”

Section: A High-level View Of New User's Preference Prediction Using mentioning

confidence: 99%

“…We show that using context similarity using neighborhood algorithm to determine the so-called neighbors of the new user, it is possible to build a simple and yet effective context-aware personalized service that can provide relevant multimedia items to first time users based on three important requirements. First, such systems should be capable of determining the new user's current context [24]. Second, it should be able to use the identified contextual information to discover users who have been in the same or similar contexts, including the content they have consumed in those contexts.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Context-Aware Personalization Using Neighborhood-Based Context Similarity

Otebolaku

Andrade

2016

Wireless Pers Commun

Self Cite

View full text Add to dashboard Cite

With the overwhelming volume of online multimedia content and increasing ubiquity of Internet-enabled mobile devices, pervasive use of the Web for content sharing and consumption has become our everyday routines. Consequently, people seeking online access to content of interest are becoming more and more frustrated. Thus, deciding which content to consume among the deluge of available alternatives becomes increasingly difficult. Contextaware personalization, having the capability to predict user's contextual preferences, has been proposed as an effective solution. However, some existing personalized systems, especially those based on collaborative filtering, rely on rating information explicitly obtained from users in consumption contexts. Therefore, these systems suffer from the socalled cold-start problem that occurs as a result of personalization systems' lack of adequate knowledge of either a new user's preferences or of a new item rating information. This happens because these new items and users have not received or provided adequate rating information respectively. In this paper, we present an analysis and design of a context-aware personalized system capable of minimizing new user cold-start problem in a mobile multimedia consumption scenario. The article emphasizes the importance of similarity between contexts of consumption based on the traditional k-nearest neighbor algorithm using Pearson Correlation model. Experimental validation, with respect to quality of personalized recommendations and user satisfaction in both contextual and non-contextual scenarios, shows that the proposed system can mitigate the effect of user-based cold-start problem.

show abstract