Future Computing and Informatics Journal最新文献

The volume of data on the social media is huge and even keeps increasing. The need for efficient processing of this extensive information resulted in increasing research interest in knowledge engineering tasks such as Opinion Summarization. This survey shows the current opinion summarization challenges for social media, then the necessary pre-summarization steps like preprocessing, features extraction, noise elimination, and handling of synonym features. Next, it covers the various approaches used in opinion summarization like Visualization, Abstractive, Aspect based, Query-focused, Real Time, Update Summarization, and highlight other Opinion Summarization approaches such as Contrastive, Concept-based, Community Detection, Domain Specific, Bilingual, Social Bookmarking, and Social Media Sampling. It covers the different datasets used in opinion summarization and future work suggested in each technique. Finally, it provides different ways for evaluating opinion summarization.

Pulsars are rare type of stars that emit radio signals that could be detected from earth. Astronomy scientists give more attention to this type of stars for many reasons. In the near past, the problem of pulsar selection was carried out manually. Recently, neural network techniques are proposed to solve the problem. In this paper, we present a novel technique to efficiently selecting pulsars. The proposed algorithm is based on the fuzzy knn classifier. Results show that, the proposed algorithm outperforms five other classifiers, including neural network classifiers, using three evaluation metrics. The proposed algorithm is evaluated on the recent HITRU 2 dataset.

Privacy problems are lethal and getting more attention than any other issue with the notion of the Internet of Things (IoT). Since IoT has many application areas including smart home, smart grids, smart healthcare system, smart and intelligent transportation and many more. Most of these applications are fueled by the resource-constrained sensor network, such as Smart healthcare system is powered by Wireless Body Area Network (WBAN) and Smart home and weather monitoring systems are fueled by Wireless Sensor Networks (WSN). In the mentioned application areas sensor node life is a very important aspect of these technologies as it explicitly effects the network life and performance. Data aggregation techniques are used to increase sensor node life by decreasing communication overhead. However, when the data is aggregated at intermediate nodes to reduce communication overhead, data privacy problems becomes more vulnerable. Different Privacy-Preserving Data Aggregation (PPDA) techniques have been proposed to ensure data privacy during data aggregation in resource-constrained sensor nodes. We provide a review and comparative analysis of the state of the art PPDA techniques in this paper. The comparative analysis is based on Computation Cost, Communication overhead, Privacy Level, resistance against malicious aggregator, sensor node life and energy consumption by the sensor node. We have studied the most recent techniques and provide in-depth analysis of the minute steps involved in these techniques. To the best of our knowledge, this survey is the most recent and comprehensive study of PPDA techniques.

Business process reengineering (BPR) is identified as one of the most important solutions for organizational improvements in all performance measures of business processes. However, high failure rates 70% is reported about using it the most important reason that caused the failure is the focus on the process itself; regardless of the surrounding environment, and the knowledge of the organization. The other reasons are due to the lack of tools to determine the causes of the inconsistencies and inefficiencies.

This paper proposes Process Reengineering Ontology-based knowledge Map Methodology (PROM) to reduce the failure ratio, solve BPR problems, and overcome their difficulties. Using an organizational ontology to show the structure and environment surrounding to organization's processes, using knowledge maps as an inference that succeeds to identify and find out the causes that lead to contradictions and inefficiencies, and using Analytical hierarchy processing to identify and prioritize processes of the business to be re-designed. Through the proposed methodology, all organizational processes are completely analyzed. Moreover, Analytical Hierarchy Processing technique is used to show the most important processes with high priority to be reengineered first then it is easy to discover any errors occurred during reengineering process through knowledge map so BPR is done successfully. Finally, Apply the proposed methodology to inventory management shows how processes reengineering are done successfully and helping the organization to achieve its objectives.

Microarray gene expression data plays a prominent role in feature selection that helps in diagnosis and treatment of a wide variety of diseases. Microarray gene expression data contains redundant feature genes of high dimensionality and smaller training and testing samples. This paper proposes a customized similarity measure using fuzzy rough quick reduct algorithm for attribute selection. Information Gain based entropy is used to reduce the dimensionality in the first stage and the proposed fuzzy rough quick reduct method that defines a customized similarity measure for selecting the minimum number of informative genes and removing the redundant genes is employed at the second stage. The proposed method is evaluated using leukemia, lung and ovarian cancer gene expression datasets on a random forest classifier. The proposed method produces 97.22%, 99.45% and 99.6% classifier accuracy on leukemia, lung and ovarian cancer gene expression datasets respectively. The research study is carried out using the R open source software package. The proposed method shows substantial improvement in the performance with respect to various statistical parameters like classification accuracy, precision, recall, f-measure and region of characteristic compared to available methods in literature.

From the human-computer interface perspectives, the challenges to be faced are related to the consideration of new, multiple interactions, and the diversity of devices. The large panel of interactions (touching, shaking, voice dictation, positioning …) and the diversification of interaction devices can be seen as a factor of flexibility albeit introducing incidental complexity. Our work is part of the field of user interface description languages. After an analysis of the scientific context of our work, this paper introduces HCIDL, a modelling language staged in a model-driven engineering approach. Among the properties related to human-computer interface, our proposition is intended for modelling multi-target, multimodal, plastic interaction interfaces using user interface description languages. By combining plasticity and multimodality, HCIDL improves usability of user interfaces through adaptive behaviour by providing end-users with an interaction-set adapted to input/output of terminals and, an optimum layout.

In this paper, we propose a novel multi-objective binary bat algorithm for simultaneous ranking and selection of keystroke dynamics features. The proposed algorithm uses the V shaped binarization function. Simulation results show that, the proposed algorithm can efficiently identify the most important features of the data set. Of the three feature classes, the key down hold time features (H-features) are proofed to be the most dominant features. Using H-features only in classification decreases the mean square error (MSE) by 2% compared to choosing all features in classification. The UD features are the second ranked features. The worst features are the DD features which represent the largest MSE when being used individually in the classification process. The results are performed using two classifiers for comparisons; the linear and the quadratic classifiers. The quadratic classifier outperforms the linear classifier with respect to the mean square error (MSE) and the average number of features selected.

Information gathering from patient by clinicians during diagnostic procedures may sometimes require some skills to adequately collect required information that will be sufficient for the procedure. A situation where this information gathering may proof difficult in when a diagnostic decision making support system (DDSS) will have to gather such information from patient before carrying out the diagnostic procedure. Research has proven that it is more challenging to ensure user or patient inputs, in their raw form, maps into the list of acceptable medical terms for diagnostic tasks. This paper therefore proposes a formalized input generating model that addresses this shortcoming through the creation of an inference process, breast cancer lexicon, rule set and natural language processing (NLP). We developed an input generation algorithm which uses the python natural language processing capability in first filtering and generation the first pre-input collection. Furthermore, this algorithm then feeds in the pre-input word collection as input into the inference engine which has in its memory the rule set and ontology-based lexicon developed. Finally, this generates a list of acceptable tokens that will be sent into the medical expert system or DDSS for the diagnosing breast cancer. This proposed model was tested on a breast cancer based DDSS earlier designed by this authors, and result shows that the inference support of this model generates additional input of about 64% compared to when the patient's input where sent in as input in is state.

Deep Learning is a new machine learning field that gained a lot of interest over the past few years. It was widely applied to several applications and proven to be a powerful machine learning tool for many of the complex problems. In this paper we used Deep Neural Network classifier which is one of the DL architectures for classifying a dataset of 66 brain MRIs into 4 classes e.g. normal, glioblastoma, sarcoma and metastatic bronchogenic carcinoma tumors. The classifier was combined with the discrete wavelet transform (DWT) the powerful feature extraction tool and principal components analysis (PCA) and the evaluation of the performance was quite good over all the performance measures.