Search

Many digital scholarly platforms allow researchers to conduct scientific discussions using their question-answer systems. Such discussions are helpful for emerging as well as established researchers in numerous prospects. In order to rely on such discussions, it is essential to effectively validate these questions and answers based on content as well as context. This will help in finding 'quality' content, improving user experience for researchers and acquiring efficient results in realistic applications such as influence analysis, topic modeling, expert finding and recommendation systems. In this research, one of the well-known scholarly platforms, ResearchGate (RG) is focused. Questions and answers of 14,000 researchers from RG are rendered. A qualitative analysis is performed to perceive various anomalies in the rendered data. To resolve these anomalies, we propose a Word2Vec based model that identifies redundancy and relevancy in questions and answers using content as well as context analysis. The proposed model is evaluated on the rendered data and the outcome suggests that the model accurately identifies redundant questions and irrelevant answers from scholarly platforms. [ABSTRACT FROM AUTHOR]

Social media is attaining popularity day by day across domains. Approximately about 40% of the world's population uses social media. Automatically these people look the influencers in social media to guide them with their decision-making in a day-to-day life. An influencer in social media is someone who can influence others. Finding the best influencer as a seed node among social media users is relevant to research. This paper proposes a methodology that can get the most suitable influencer list based on context analysis. A contextual analysis analyzes a text in different media that helps us assess that text within the context. The hybrid way of using the graph networks and the machine learning algorithms like LSTM and Naive Bayes together as suggested in this article helps us assess that text and find the influencer significantly better than the existing approaches. We prove that the hybrid model can produce more accurate results in such a way that we only get the users who can positively influencer other users. Based on our result, we are sure that this hybrid model can produce more accurate results for funding the best set of influencers. Along with that, the combined use of LSTM and Naive Bayes can produce better results. [ABSTRACT FROM AUTHOR]

Traditional collaborative filtering based recommender systems deal with the two-dimensional user-item rating matrix where users have rated a set of items into the system. Although traditional recommender systems are widely adopted but they are unable to generate effective recommendations in case of multi-dimensionality i.e. multi-criteria ratings, contextual information, side information etc. The curse of dimensionality is the major issue in the recommendation systems. Therefore, in this paper, we proposed a clustering approach to incorporate multi-criteria ratings into traditional recommender systems effectively. Furthermore, we compute the intra-cluster user similarities using a Mahalanobis distance method in order to make more accurate recommendations and compared the proposed approach with the traditional collaborative filtering based method using Yahoo! Movies dataset. [ABSTRACT FROM AUTHOR]

Recommendation Systems have gained huge importance in recent years as they proffer most admissible choices explored out of large data sets and are analogous to the user’s interests. But the notion of context – that describes the user, item or how users interact with the recommendation system has become one of the most significant factors that has increased the potency of recommendation systems in distinct domains such as e-commerce, movies, music, tourism etc. The main key issues in context aware recommendations are how to utilize the contextual information efficiently. So to utilize this information efficiently, an innovative cluster based context-aware post-filtering approach has been propounded that implicitly segregates the users into clusters on the basis of high performing contextual dimensions; combine the clusters on the basis of contextual similarity and then re-rank the items according to the probability of relevance in a cluster. The proposed approach is validated on movies dataset consisting of different contextual dimensions. The empirical analysis validates that the propounded approach surpass various state-of-the-art approaches. [ABSTRACT FROM AUTHOR]

The objective of this study is to assess the capability and priority for Enterprise Architecture (EA) implementation in the context of Malaysian public sector agency. This study uses the Analytic Hierarchy Process (AHP) based on the judgments and opinions of EA team members from three different agencies at federal level. There are 27 assessment criteria which are grouped into six categories, Internal Process, Learning and Growth, Authority Support, Cost, Technology and Talent Management. The top capability of all cases is Internal Process and the top criteria for priority assessment is Authority Support. From the findings it can be concluded that AHP is a useful and cost effective method to assess, priorities and plan for EA implementation. Addressing the capability and priority criteria will ensure the optimum EA is implemented thus it shall reduce the risk of EA implementation failure. [ABSTRACT FROM AUTHOR]

The enterprise applications consist of a set of applications that make up the software ecosystem. These can be developed in-house or acquired from third party companies. Generally these applications are developed in order to meet a demands/specific business need of the company without worrying of interacting with another existing system. The area of Enterprise Application Integration (EAI) is able to provide all techniques and tools in order to integrate heterogeneous applications software ecosystem of companies. The Guaraná DSL is a language that enables designing conceptual models of integration solutions using a concrete graphical and intuitive syntax. Integrating the applications is not a trivial task and solution development involves addition of costs (time and resources), risks such as the appearance of bugs that most often are observed only after the implementation of the conceptual model. In this paper we propose a simulation model using Petri Nets that enable the analysis of the behavior of an integration solution developed to a real-world integration problem at Unijuí University. [ABSTRACT FROM AUTHOR]

Phenotypic information derived from visual characteristics of colorectal cancer (CRC) is routinely used for diagnosis and recommendations for treatment. Previously published studies show that the ratio of tissue types within CRC is prognostic. Such studies generate large amounts of data, combining expert classifications with x-y coordinates, which has previously been used to train image analysis algorithms. This paper describes extensions to algorithms employed in previously published work, using pixel clustering as a pre-processing step before normalised cuts in order to reduce the size of the graph for unsupervised segmentation. Image segments are processed for features and given a candidate classification which is weighted by neighbouring segment classes. Global slide features are incorporated to mitigate inconsistencies in overall appearance caused by histological and biological differences. The proposed algorithm increases agreement with the ground truth from 75% to 79% on a dataset of 7,159 images across 157 digital slides. [ABSTRACT FROM AUTHOR]

For a collaborative assistive device, human intent recognition (IR) is one of the first and foremost requirements. Formalizing the complex process of human IR in a compact yet expressive mathematical model holds promise. We put forward a Hierarchical Finite State Machine (H-FSM) for human IR within a generalized framework for collaborative assistive devices. Visual and contextual observations drive the H-FSM to different levels of granularity. [ABSTRACT FROM AUTHOR]

Weather parameters are critical parameter which is to be considered for understanding atmospheric variation. The aim of this paper is to see how the parameter values fluctuates for three decades. Three decade information is isolated into two and K-means machine learning calculation is utilized for bunching the divided data. Then how the information focuses are conveyed around the centroid of every sectioned data group is confirmed. At long last the group for initial one and half decade is compared with the second one and half decade to know how the information focuses are strayed in dynamic year. This gathering is validated with silhouette value. [ABSTRACT FROM AUTHOR]

Part of Speech (POS) Tagging is a challenging task to identify the meaning of each word in a sentence. This paper shows the task of identifying each word in an odia sentence using the technique of Support Vector Machine. The POS Tagger is developed using a very small tagset of five tags. Various features sets are taken for different contextual information is helpful in predicting the POS classes. An Odia corpus of 10,000 words has taken and tested it very carefully. The previous POS Tagger was done using Artificial Neural Network (ANN) had given the accuracy of 81%. But this SVM based POS Tagger for Odia gives the result with an accuracy of 82%. It is very helpful to use in many field of natural language process. The result of this system compares with POS tagger using ANN which was previously done. [ABSTRACT FROM AUTHOR]

Abstract: This work aims to mathematically formalize the notion of context, with the purpose of allowing contextual decision-making in order to improve performance in dynamic data driven classification systems. We present definitions for both intrinsic context, i.e. factors which directly affect sensor measurements for a given event, as well as extrinsic context, i.e. factors which do not affect the sensor measurements directly, but do affect the interpretation of collected data. Supervised and unsupervised modeling techniques to derive context and context labels from sensor data are formulated. Here, supervised modeling incorporates the a priori known factors affecting the sensing modalities, while unsupervised modeling autonomously discovers the structure of those factors in sensor data. Context-aware event classification algorithms are developed by adapting the classification boundaries, dependent on the current operational context. Improvements in context-aware classification have been quantified and validated in an unattended sensor-fence application for US Border Monitoring. Field data, collected with seismic sensors on different ground types, are analyzed in order to classify two types of walking across the border, namely, normal and stealthy. The classification is shown to be strongly dependent on the context (specifically, soil type: gravel or moist soil). [Copyright &y& Elsevier]

Abstract: Previously, the reference place in intra-team interactions has been defined as the collection of references that team members share to increase consensus and harmony. Psychological studies have characterized the reference place based on verbal communication. These studies also have highlighted the need for being concerned about the other person, which is captured in the Japanese concept of Omoiyari. The current study has investigated an enhancement of such communication by exchanging images. For this purpose, a system for communicating images as well as text constrained by Omoiyari-type concerns has been developed. The system formed the basis of a study where a stranger as well as friends of the participants communicated liked and disliked images. Participants responded to that image presentation with text messages. The dependent variable was defined as the increase in harmony assessment after image presentation and was expressed in values of a Likert scale. The results showed a significant difference between images sent by strangers and by friends (F1,11 = 30.31, p < 0.005) and between liked and disliked images (F1,11 = 14.19, p < 0.01). These differences indicate a confirmatory bias, as more information is available about friends than strangers. The confirmatory bias hypothesis was further supported by a qualitative analysis of participants’ responses to images. The result suggests that exchanging images has a positive effect on relationships, if the partners have contextual knowledge available and if they use this knowledge for an Omoiyari-type approach in selecting images as opposed to using images for self-presentation. [Copyright &y& Elsevier]