A Cognitive Query Model for Arabic based on probabilistic associative morpho-phonetic Sub-Networks

This paper is discussing some novel aspects related to formalizing a Cognitive Query Model for Arabic based on constructing query associative morpho-phonetic Sub-Networks in the context of Arabic query analysis and expansion. As Humans tend to use limited number of words with possibly incomplete and ambiguous representation for requesting information, predicting the intended information conveyed in a query keywords might affect an inter-cognitive communication dramatically. Based on Associative Probabilistic Bi-directional Root-Pattern Relations introduced in APRoPAT Statistical Language Model, a cognitively motivated representation for query semantic network construction is proposed. This Model attempts to predicting the most plausible intended query information by constructing a morpho-phonetic cognitive Sub-Network based on the query terms and instantiation of the most probable query root-pattern phonetic vectors within the global Associative Network expressed by the APRoPAT Model.