Abstract The Bayesian model for prediction problems requires setting up the prior/hyper prior structures that go through the process of integration. However, these formulated integrals are not tractable analytically. Moreover, application of Markov Chain Monte Carlo (MCMC) methods to solve these integralsare slow in nature, especially if the parameter space is high dimensional. The key idea behind the Bayesian inference is to... The Bayesian model for prediction problems requires setting up the prior/hyper prior structures that go through the process of integration. However, these formulated integrals are not tractable analytically. Moreover, application of Markov Chain Monte Carlo (MCMC) methods to solve these integralsare slow in nature, especially if the parameter space is high dimensional. The key idea behind the Bayesian inference is to marginalize over unknown parameters, rather than make point estimation. This technique avoids severe over-fitting problems and allows direct model comparison. In this paper, we presented a model using variational Bayesian inference for identification and interpretation of Non Standard Words in Bengali news corpus. The variational methods extend the practicality of Bayesian inference to complex Bayesian models and “medium sized” data sets. The Variational Bayesian inference aims to approximate posterior distribution by a simpler distribution for which marginalization is tractable. Read More Read Less

Abstract Corpus based stemming has been devised to develop stemmers targeting language independent environment. These stemmers are applicable to all languages based on Latin script. In the present work, we exploit the English corpus for coded words of Devanagari script. We use the technique of Romanization and the stemmer is being tested over 100 randomly chosen Hindi words. We show that this... Corpus based stemming has been devised to develop stemmers targeting language independent environment. These stemmers are applicable to all languages based on Latin script. In the present work, we exploit the English corpus for coded words of Devanagari script. We use the technique of Romanization and the stemmer is being tested over 100 randomly chosen Hindi words. We show that this approach has a direct application to the standardization of regional languages. For instance, we standardize the Kumauni language. Read More Read Less

Abstract Natural Language Constraints have a vital role in Business Organization. The main problem is scope ambiguity when these NL constraints are translated into formal languages. Human beings can understand the context in which these constraints are defined but it is most difficult for a machine to understand the exact meanings of these constraints in their context and this leads to crash... Natural Language Constraints have a vital role in Business Organization. The main problem is scope ambiguity when these NL constraints are translated into formal languages. Human beings can understand the context in which these constraints are defined but it is most difficult for a machine to understand the exact meanings of these constraints in their context and this leads to crash the Business System. Therefore before translating these business constraints to OCL, the scope ambiguities should be resolved for correct translation of NL to OCL. For this purpose a new technique is proposed for handling the scope of logical operators used in NL constraints by using the Markov Logic. The subject and scope knowledge and Markov logic integrates into natural language processing. The subject knowledge has been deployed as context knowledge and scope knowledge was acquired from the business constraints. Markov logic was applied to NL constraints for selecting the most possible meaning of an ambiguous NL constraint based on the context. The presented work shows that by handling the identified cases of scope ambiguities of logical operators, we can take correct translation of business constraints into formal specifications. Read More Read Less