Devendra K. Tayal

Measuring context-meaning for open class words in Hindi language

Word sense disambiguation (WSD), the task of identifying the intended sense of words has been a growing research area in the field of natural language processing. In this paper, the authors focused on word sense disambiguation for Hindi language using graph connectivity measures and Hindi WordNet[1]. To construct the graph for the sentence each sense of the ambiguous word is taken as a source node and all the paths which connects the sense to other words present in the sentence are added. The importance of nodes in the constructed graph are identified using node neighbor based measures (various centrality) and graph clustering based measures (denseness, graph randomness, edge density). The proposed method disambiguates all open class words (noun, verb, adjective, adverb) and disambiguates all the words present in the sentence simultaneously.

On reverse engineering an object-oriented code into UML class diagrams incorporating extensible mechanisms

Reverse engineering is the key idea for reconstruction of any existing system. In this paper, we propose an algorithm to reverse engineer an object-oriented code into Unified Modeling Language (UML) class diagram. Our algorithm is very general in nature and can be applied to any object-oriented code irrespective of the object-oriented programming language. In our paper we consider an object-oriented pseudocode similar to C++ to implement our algorithm. Some of the researchers have dealt in the past the problem of reverse engineering an object-oriented code to UML class diagrams. However, none of these researchers have treated all the constructs available in UML class diagrams. Unlike the previously done work on reverse engineering into UML, our algorithm generates rules for a complete set of constructs available in UML class diagrams. It includes classes, relationships, objects, attributes, operations, inheritance, associations, interfaces & other extensible mechanisms also. This algorithm can be viewed as a solution to reverse engineer any available object-oriented software. An application for the implementation of above said rules using C++ code is also included in the paper. We thoroughly compare our work with the similar type of earlier work in this area and uncover the deficiencies in these previous available works. Moreover our motive in this paper is to prepare rules to reverse engineer C++ code into UML class diagrams and not to generate any tools.

Supervised Metaphor Detection using Conditional Random Fields

In this paper, we propose a novel approach for supervised classification of linguistic metaphors in an open domain text using Conditional Random Fields (CRF). We analyze CRF based classification model for metaphor detection using syntactic, conceptual, affective, and word embeddings based features which are extracted from MRC Psycholinguistic Database (MRCPD) and WordNet-Affect. We use word embeddings given by Huang et al. to capture information such as coherence and analogy between words. To tackle the bottleneck of limited coverage of psychological features in MRCPD, we employ synonymy relations from WordNet ® . A comparison of our approach with previous approaches shows the efficacy of CRF classifier in detecting metaphors. The experiments conducted on VU Amsterdam metaphor corpus provides an accuracy of more than 92% and Fmeasure of approximately 78%. Results shows that inclusion of conceptual features improves the recall by 5% whereas affective features do not have any major impact on metaphor detection in open text.

MetaSurfer: a new metasearch engine based on FAHP and modified EOWA operator

A meta-search engine (MSE) is a system that supports unified access to multiple existing web search engines by querying them simultaneously, thereby widening coverage of the World Wide Web. Result aggregation techniques are used in an MSE to combine the results from multiple underlying web search engines and present a single consolidated result list to the user. In this paper, we propose a new method for optimization of MSE results using fuzzy analytical hierarchy process (FAHP) and modified extended ordered weighted averaging (EOWA) operator. In our method, FAHP is used for pair-wise comparison of documents in underlying search engines. We assign fuzzy importance degrees to search engines which are represented in triangular fuzzy numbers, and subsequently defuzzified using center of gravity method. Center of gravity method is found to be a better defuzzifier operator in the literature as compared to max membership method which is an important constituent of EOWA operator. The modified EOWA operator is therefore proposed to aggregate the scores of individual search engine documents so as to obtain the final ranking of the documents. Lastly, we compute the effectiveness of the proposed MSE viz. MetaSurfer by calculating the classical Precision. Since we have proposed a new method which reorders documents such that a more relevant document occupies a higher rank, therefore we have defined a new criterion to calculate effectiveness called weighted precision. The precision and weighted precision values are compared with those of existing and popular MSEs. The results show that MetaSurfer performs better than existing MSEs.

Automatically incorporating context meaning for query expansion using graph connectivity measures

In order to improve the retrieval performance, the query is reformulated by the process of Query expansion (QE). Most of the existing query expansion techniques do not consider the context of the terms present in the user’s query which can result in low precision and recall. Through this paper, the query consisting of ambiguous terms (polysemy words) is expanded by selecting the terms, which are in close proximity to the query terms while context meaning of the terms is automatically incorporated. The basis of this query expansion method is to investigate the role of graph structure (which is being created for the query) and determining the importance of each node in the graph using WordNet. The relevant nodes representing word senses are identified from the graph and can be chosen as additional terms to be added to the query for improving the retrieval of web pages. The experiments, conducted on data sets of ambiguous queries show that proposed approach outperforms other query expansion methodologies by enhancing precision and recall.

ClusFuDE: Forecasting low dimensional numerical data using an improved method based on automatic clustering, fuzzy relationships and differential evolution

Abstract In this paper, a novel hybrid model for forecasting low dimensional numerical data is proposed which is named as ClusFuDE. The proposed method uses an improved automatic clustering approach for clustering the historical numerical data. Further fuzzy logical relationships are used to forecast the approximate values which are then defuzzified to calculate the exact forecasted values of the data. The fuzzy logical relationships are useful in modelling the fuzzy relations and help in forecasting the fuzzy time series data in a very simplified manner. The forecasted sub-optimal candidate solutions are optimized using Differential Evolution. The Differential Evolution method uses a dynamic differential crossover rate ( C r i ) for the i th solution, for identifying and discarding suboptimal candidate solutions in early stages of the iterative run. This makes the method more suitable for iterative modification of candidate solutions by using differential mutation and crossover, and suitable for global search. The proposed method is applied for forecasting, the year wise enrollments of the University of Alabama, Lahi (crop) production, monthly amount of outpatients visit in a hospital, inventory demand and population of India from years 1930–2000 and the results are consistent. We have compared our method with the recent forecasting methods available in literature and the proposed method outperforms all the existing methods in the literature. The accuracy of the proposed method is computed by calculating the Mean square error (MSE) and Mean Absolute percentage error (MAPE). The proposed method provides the lowest MSE and MAPE when compared to all other methods available in the literature.

Identifying Metaphors Using Fuzzy Conceptual Features

Metaphor comprehension is a challenging problem which equally intrigues researchers in linguistics as well as those working in the domain of cognition. The use of psychological features such as Imageability and Concreteness has been shown to be effective in identifying metaphors. However, there is a certain degree of vagueness and blurring boundaries between the sub-concepts of these features that has hitherto been largely ignored. In this paper, we tackle this issue of vagueness by proposing a fuzzy framework for metaphor detection whereby linguistic variables are employed to express psychological features. We develop a Mamdani Model to extract fuzzy classification rules for detecting metaphors in a text. The results of experiments conducted over a dataset of nominal metaphors reveal encouraging results with an F-score of more than 79%.

A Novel Method for Test Path Prioritization using Centrality Measures

Software testingisanessentialstageof thesoftwaredevelopment lifecyclewhich helpsinproducingbugfreesoftware.Thispaperintroducesastrategytocreatetest dataconsequentlyfromthebeginningoftestinformationwhichistestedagainstthe Programundertest(PUT)foramplenesscriteria.Initiallythisprocessproducestest information set arbitrarilywhere auniqueapproach for the test pathprioritization process is presented that uses the concept of centrality measures. The proposed algorithmcomputestheimportanceofeachnodeinthetestpathsbyusingvarious centralitymeasuresand thus identifies thesignificanceofeachpath.Furthermore, theproposedmethodologyshowsthemaximumnumberofpotentialnodesthatare covered using a less number of prioritized paths. This paper tests the created test informationagainsttheproducttocheckitssufficiency. KeywORdS Betweenness, Centrality Measures, Closeness, Degree, Path Prioritization, Software Testing

Big Data’s Biggest Problem: Load Imbalancing

Uneven distribution of data between the nodes causes the data skewness problem. Due to this problem, various problems occur during the processing. So, this paper presents the brief analysis of the existing techniques related to load imbalancing with their pros and cons. Also, types of data skewness have been discussed in this paper.

Text Summarization Using WordNet Graph Based Sentence Ranking

Text summarization refers to the task of generating a summary from a given document that tries to replicate the most significant information of the original document. A number of techniques are available in the literature regarding the same and sentence ranking is one of them. This paper proposes a novel method for text summarization using WordNet graph based sentence ranking. The proposed method utilizes the degree, betweenness, and closeness centrality measures. This paper also analyzes the current research work going on in text summarization and sentence ranking. Web of Science (WoS) is used as the data source for the same. The co-occurrence of all the keywords in the research papers pertaining to sentence ranking is also visualized.