Abstract Case-Based Reasoning (CBR) often shows significant promise for improving the effectiveness of decision support systems. Previous research in decision making has proposed the use of a CBR approach to accumulate, organize, preserve, link and share diverse knowledge coming from past experiences. However, existing CBR systems lack semantic understanding, which is important for intelligent knowledge retrieval in decision support systems. In... Case-Based Reasoning (CBR) often shows significant promise for improving the effectiveness of decision support systems. Previous research in decision making has proposed the use of a CBR approach to accumulate, organize, preserve, link and share diverse knowledge coming from past experiences. However, existing CBR systems lack semantic understanding, which is important for intelligent knowledge retrieval in decision support systems. In order to overcome the limitation of existing CBR and develop an intelligent CBR system which can not only carry out data matching retrieval, but also perform semantic associated data access, and improve the traditional keyword-based search, this paper integrates ontology technology into a CBR system and proposes to combine semantic retrieval method and numerical measurement in case retrieval. Ontology technology is an ideal selection for realizing our system thanks to the good semantic understanding offered by ontology. The resulting ontology based CBR system is applied in fault diagnosis of industrial machines, a semi-structured decision-making environment involving multiple attributes. The case illustrates the use of proposed semantic CBR system, and shows the feasibility of our approach and the benefit of the ontology support. Read More Read Less

Abstract We develop an extended version of spell checker and corrector to check non-word errors in social media datasets, which will be used in message filtering systems especially for cyberbullying detection. We use the dictionary techniques to check words, twelve-word spell error checking and correction approaches to correct the non-word errors, and n-gram and Levenshtein distance to select the most suitable word among corrected words. If there is more... We develop an extended version of spell checker and corrector to check non-word errors in social media datasets, which will be used in message filtering systems especially for cyberbullying detection. We use the dictionary techniques to check words, twelve-word spell error checking and correction approaches to correct the non-word errors, and n-gram and Levenshtein distance to select the most suitable word among corrected words. If there is more than one corrected word we get from each approach, we use n-gram techniques to choose the corrected and reasonable word from the words in n-gram database. When we used the Levenshtein distance in our previous work, we found that it selected the first corrected word and it was not a reasonable one in some sentences. Therefore, we use the n-gram database in this paper. Read More Read Less

Abstract Community Question-Answering(CQA) sites are the major platform where posts are generated by peers in the form of questions and answers for information seeking in online environments. In general, multiple posts are created by different users on a particular topic or subject. Large number of posts raises the difficulties in information management of these sites. A number of approaches are suggested in recent research work for efficient management of... Community Question-Answering(CQA) sites are the major platform where posts are generated by peers in the form of questions and answers for information seeking in online environments. In general, multiple posts are created by different users on a particular topic or subject. Large number of posts raises the difficulties in information management of these sites. A number of approaches are suggested in recent research work for efficient management of data for CQA sites. Many of the existing approaches have suggested use of clustering techniques for managing the CQA sites, but ignored the tagging data (user tags) of the posts. In this paper, an improved spectral clustering technique is derived based on similarity measures for text processing (SMTP) and utilized for clustering the posts considering the tagging data available on CQA sites. A specialized data structure, namely, folksonomy is developed for clustering using the relationship between tags, posts and users.The proposed method is developed in two stages. In first stage, the folksonomy relation is created and post similarity graph is built with the help of tag frequency-inverse post frequency. In the second stage, the spectral clustering algorithm is applied to the post similarity matrix to group the similar posts. The post clusters are generated as the output of the proposed algorithm where the post information management can be made with the help of user tags. The experimental results show that the improvedspectral clustering algorithm outperforms the other considered clustering algorithms with a huge margin and it will be helpful for information management of the CQA sites. The improved spectral clustering algorithm will be useful for post pre-processing, faster information retrieval, duplicate posts identification and posts management from clusters. Read More Read Less

Abstract The aim of this paper is to improve the search query performance of the biomedical literature by expanding the queries with most significant terms. Methods: In this article, an enhanced BM25 mathematical approach is proposed to retrieve the most query relevant literature from clustered Biomedical literature bank with query expansion from MeSH. The clustered biomedical topics are analyzed with different pre-processing... The aim of this paper is to improve the search query performance of the biomedical literature by expanding the queries with most significant terms. Methods: In this article, an enhanced BM25 mathematical approach is proposed to retrieve the most query relevant literature from clustered Biomedical literature bank with query expansion from MeSH. The clustered biomedical topics are analyzed with different pre-processing methods and term-weighting functions and found the best values for the tuning parameters K1, b, K3 and the right combination of pre-processing and term weighting functions for improving the query performance in terms of Average Precision, Mean Average Precision and R-precision. Read More Read Less