Search results

We describe an intelligent video categorization engine (IVCE) that uses the learning capability of artificial neural networks (ANNs) to classify suitably preprocessed video segments into a predefined number of semantically meaningful events (categories).

We provide a survey of existing techniques that have been proposed, either directly or indirectly, towards achieving intelligent video categorization. We also compare the performance of two popular ANNs: Kohonen's self‐organizing map (SOM) and fuzzy adaptive resonance theory (Fuzzy ART). In particular, the ANNs are trained offline to form the necessary knowledge base prior to online categorization.

Experimental results show that accurate categorization can be achieved near instantaneously.

The main limitation of this research is the need for a finite set of predefined categories. Further research should focus on generalization of such techniques.

Machine understanding of video footage has tremendous potential for three reasons. First, it enables interactive broadcast of video. Second, it allows unequal error protection for different video shots/segments during transmission to make better use of limited channel resources. Third, it provides intuitive indexing and retrieval for video‐on‐demand applications.

Service-oriented architecture is an emerging software architecture, in which web service (WS) plays a crucial role. In this architecture, the task of WS composition and verification is required when handling complex requirement of services from users. When the number of WS becomes very huge in practice, the complexity of the composition and verification is also correspondingly high. In this paper, the authors aim to propose a logic-based clustering approach to solve this problem by separating the original repository of WS into clusters. Moreover, they also propose a so-called quality-controlled clustering approach to ensure the quality of generated clusters in a reasonable execution time.

The approach represents WSs as logical formulas on which the authors conduct the clustering task. They also combine two most popular clustering approaches of hierarchical agglomerative clustering (HAC) and k-means to ensure the quality of generated clusters.

This logic-based clustering approach really helps to increase the performance of the WS composition and verification significantly. Furthermore, the logic-based approach helps us to maintain the soundness and completeness of the composition solution. Eventually, the quality-controlled strategy can ensure the quality of generated clusters in low complexity time.

The work discussed in this paper is just implemented as a research tool known as WSCOVER. More work is needed to make it a practical and usable system for real life applications.

In this paper, the authors propose a logic-based paradigm to represent and cluster WSs. Moreover, they also propose an approach of quality-controlled clustering which combines and takes advantages of two most popular clustering approaches of HAC and k-means.

In particular, this paper aims to systematically analyze a few prominent wireless sensor network (WSN) clustering routing protocols and compare these different approaches according to the taxonomy and several significant metrics.

In this paper, the authors have summarized recent research results on data routing in sensor networks and classified the approaches into four main categories, namely, data-centric, hierarchical, location-based and quality of service (QoS)-aware, and the authors have discussed the effect of node placement strategies on the operation and performance of WSNs.

Performance-controlled planned networks, where placement and routing must be intertwined and everything from delays to throughput to energy requirements is well-defined and relevant, is an interesting subject of current and future research. Real-time, deadline guarantees and their relationship with routing, mac-layer, duty-cycles and other protocol stack issues are interesting issues that would benefit from further research.

Tens of thousands of news articles are posted online each day, covering topics from politics to science to current events. To better cope with this overwhelming volume of information, RSS (news) feeds are used to categorize newly posted articles. Nonetheless, most RSS users must filter through many articles within the same or different RSS feeds to locate articles pertaining to their particular interests. Due to the large number of news articles in individual RSS feeds, there is a need for further organizing articles to aid users in locating non‐redundant, informative, and related articles of interest quickly. This paper aims to address these issues.

The paper presents a novel approach which uses the word‐correlation factors in a fuzzy set information retrieval model to: filter out redundant news articles from RSS feeds; shed less‐informative articles from the non‐redundant ones; and cluster the remaining informative articles according to the fuzzy equivalence classes on the news articles.

The clustering approach requires little overhead or computational costs, and experimental results have shown that it outperforms other existing, well‐known clustering approaches.

The clustering approach as proposed in this paper applies only to RSS news articles; however, it can be extended to other application domains.

The developed clustering tool is highly efficient and effective in filtering and classifying RSS news articles and does not employ any labor‐intensive user‐feedback strategy. Therefore, it can be implemented in real‐world RSS feeds to aid users in locating RSS news articles of interest.

The need of tools for content analysis, information extraction and retrieval of multimedia objects in their native form is strongly emphasized into the judicial domain: digital videos represent a fundamental informative source of events occurring during judicial proceedings that should be stored, organized and retrieved in short time and with low cost. This paper seeks to address these issues.

In this context the JUMAS system, stem from the homonymous European Project (www.jumasproject.eu), takes up the challenge of exploiting semantics and machine learning techniques towards a better usability of multimedia judicial folders.

In this paper one of the most challenging issues addressed by the JUMAS project is described: extracting meaningful abstracts of given judicial debates in order to efficiently access salient contents. In particular, the authors present an ontology enhanced multimedia summarization environment able to derive a synthetic representation of judicial media contents by a limited loss of meaningful information while overcoming the information overload problem.

The adoption of ontology‐based query expansion has made it possible to improve the performance of multimedia summarization algorithms with respect to the traditional approaches based on statistics. The effectiveness of the proposed approach has been evaluated on real media contents, highlighting a good potential for extracting key events in the challenging area of judicial proceedings.

Nowadays, the event detection is so important in gathering news from social media. Indeed, it is widely employed by journalists to generate early alerts of reported stories. In order to incorporate available data on social media into a news story, journalists must manually process, compile and verify the news content within a very short time span. Despite its utility and importance, this process is time-consuming and labor-intensive for media organizations. Because of the afore-mentioned reason and as social media provides an essential source of data used as a support for professional journalists, the purpose of this paper is to propose the citizen clustering technique which allows the community of journalists and media professionals to document news during crises.

The authors develop, in this study, an approach for natural hazard events news detection and danger citizen’ groups clustering based on three major steps. In the first stage, the authors present a pipeline of several natural language processing tasks: event trigger detection, applied to recuperate potential event triggers; named entity recognition, used for the detection and recognition of event participants related to the extracted event triggers; and, ultimately, a dependency analysis between all the extracted data. Analyzing the ambiguity and the vagueness of similarity of news plays a key role in event detection. This issue was ignored in traditional event detection techniques. To this end, in the second step of our approach, the authors apply fuzzy sets techniques on these extracted events to enhance the clustering quality and remove the vagueness of the extracted information. Then, the defined degree of citizens’ danger is injected as input to the introduced citizens clustering method in order to detect citizens’ communities with close disaster degrees.

Empirical results indicate that homogeneous and compact citizen’ clusters can be detected using the suggested event detection method. It can also be observed that event news can be analyzed efficiently using the fuzzy theory. In addition, the proposed visualization process plays a crucial role in data journalism, as it is used to analyze event news, as well as in the final presentation of detected danger citizens’ clusters.

The introduced citizens clustering method is profitable for journalists and editors to better judge the veracity of social media content, navigate the overwhelming, identify eyewitnesses and contextualize the event. The empirical analysis results illustrate the efficiency of the developed method for both real and artificial networks.

Organizational psychologists and human resource management (HRM) practitioners often have to select the “right fit” candidate by manually scouting data from various sources including job portals and social media. Given the constant pressure to lower the recruitment costs and the time taken to extend an offer to the right talent, the HR function has to inevitably adopt data analytics and machine learning for employee selection. This paper aims to propose the “Quality of Hire” concept for employee selection using the person-environment (P-E) fit theory and machine learning.

The authors demonstrate the aforementioned concept using a clustering algorithm, namely, partition around mediod (PAM). Based on a curated data set published by the IBM, the authors examine the dimensions of different P-E fits and determine how these dimensions can lead to selection of the “right fit” candidate by evaluating the outcome of PAM.

The authors propose a multi-level fit model rooted in the P-E theory, which can improve the quality of hire for an organization.

Theoretically, the authors contribute in the domain of quality of hire using a multi-level fit approach based on the P-E theory. Methodologically, the authors contribute in expanding the HR analytics landscape by implementing PAM algorithm in employee selection.

The proposed work is expected to present a useful case on the application of machine learning for practitioners in organizational psychology, HRM and data science.

The purpose of this paper is to propose a new methodology that handles the issue of the dynamic behavior of customers over time.

A new methodology is presented based on time series clustering to extract dominant behavioral patterns of customers over time. This methodology is implemented using bank customers’ transactions data which are in the form of time series data. The data include the recency (R), frequency (F) and monetary (M) attributes of businesses that are using the point-of-sale (POS) data of a bank. This data were obtained from the data analysis department of the bank.

After carrying out an empirical study on the acquired transaction data of 2,531 business customers that are using POS devices of the bank, the dominant trends of behavior are discovered using the proposed methodology. The obtained trends were analyzed from the marketing viewpoint. Based on the analysis of the monetary attribute, customers were divided into four main segments, including high-value growing customers, middle-value growing customers, prone to churn and churners. For each resulted group of customers with a distinctive trend, effective and practical marketing recommendations were devised to improve the bank relationship with that group. The prone-to-churn segment contains most of the customers; therefore, the bank should conduct interesting promotions to retain this segment.

The discovered trends of customer behavior and proposed marketing recommendations can be helpful for banks in devising segment-specific marketing strategies as they illustrate the dynamic behavior of customers over time. The obtained trends are visualized so that they can be easily interpreted and used by banks. This paper contributes to the literature on customer relationship management (CRM) as the proposed methodology can be effectively applied to different businesses to reveal trends in customer behavior.

In the current business condition, customer behavior is changing continually over time and customers are churning due to the reduced switching costs. Therefore, choosing an effective customer segmentation methodology which can consider the dynamic behaviors of customers is essential for every business. This paper proposes a new methodology to capture customer dynamic behavior using time series clustering on time-ordered data. This is an improvement over previous studies, in which static segmentation approaches have often been adopted. To the best of the authors’ knowledge, this is the first study that combines the recency, frequency, and monetary model and time series clustering to reveal trends in customer behavior.

The fabulous results of convolution neural networks in image-related tasks attracted attention of text mining, sentiment analysis and other text analysis researchers. It is, however, difficult to find enough data for feeding such networks, optimize their parameters, and make the right design choices when constructing network architectures. The purpose of this paper is to present the creation steps of two big data sets of song emotions. The authors also explore usage of convolution and max-pooling neural layers on song lyrics, product and movie review text data sets. Three variants of a simple and flexible neural network architecture are also compared.

The intention was to spot any important patterns that can serve as guidelines for parameter optimization of similar models. The authors also wanted to identify architecture design choices which lead to high performing sentiment analysis models. To this end, the authors conducted a series of experiments with neural architectures of various configurations.

The results indicate that parallel convolutions of filter lengths up to 3 are usually enough for capturing relevant text features. Also, max-pooling region size should be adapted to the length of text documents for producing the best feature maps.

Top results the authors got are obtained with feature maps of lengths 6–18. An improvement on future neural network models for sentiment analysis could be generating sentiment polarity prediction of documents using aggregation of predictions on smaller excerpt of the entire text.

Among the growing number of data mining (DM) techniques, outlier detection has gained importance in many applications and also attracted much attention in recent times. In the past, outlier detection researched papers appeared in a safety care that can view as searching for the needles in the haystack. However, outliers are not always erroneous. Therefore, the purpose of this paper is to investigate the role of outliers in healthcare services in general and patient safety care, in particular.

It is a combined DM (clustering and the nearest neighbor) technique for outliers’ detection, which provides a clear understanding and meaningful insights to visualize the data behaviors for healthcare safety. The outcomes or the knowledge implicit is vitally essential to a proper clinical decision-making process. The method is important to the semantic, and the novel tactic of patients’ events and situations prove that play a significant role in the process of patient care safety and medications.

The outcomes of the paper is discussing a novel and integrated methodology, which can be inferring for different biological data analysis. It is discussed as integrated DM techniques to optimize its performance in the field of health and medical science. It is an integrated method of outliers detection that can be extending for searching valuable information and knowledge implicit based on selected patient factors. Based on these facts, outliers are detected as clusters and point events, and novel ideas proposed to empower clinical services in consideration of customers’ satisfactions. It is also essential to be a baseline for further healthcare strategic development and research works.

This paper mainly focussed on outliers detections. Outlier isolation that are essential to investigate the reason how it happened and communications how to mitigate it did not touch. Therefore, the research can be extended more about the hierarchy of patient problems.

DM is a dynamic and successful gateway for discovering useful knowledge for enhancing healthcare performances and patient safety. Clinical data based outlier detection is a basic task to achieve healthcare strategy. Therefore, in this paper, the authors focussed on combined DM techniques for a deep analysis of clinical data, which provide an optimal level of clinical decision-making processes. Proper clinical decisions can obtain in terms of attributes selections that important to know the influential factors or parameters of healthcare services. Therefore, using integrated clustering and nearest neighbors techniques give more acceptable searched such complex data outliers, which could be fundamental to further analysis of healthcare and patient safety situational analysis.