Hire Ashkan G. - NLP Engineer

* Design and build a naive Bayes classifier for predicting the price category of question. Build a price model to cluster the price between 5 different categories and combine it with category of question to train a multi-class classifier to increase the accuracy of the classifier. * Patent, design and implement an innovative application to follow interesting questions, topics and categories base on user behavior and interaction with the system. Implement topic extraction algorithm to identify important topics for each category and used extracted topics to build auto-topic suggestion when user search for topics. Build a phrasal query search interface to find similar questions and relevant questions using Solr. * Work on improving the accuracy of a question categorizer system. Implemented a naive bayes classifier for more than 1000 categories. Create n-gram features and use Medical, Vet and Car Ontology to enhance the accuracy of feature probability. Research and implemented a probability confidence model to convert the basysian probability to a more meaningful confidence score. Improved the accuracy of classifier for set of target categories by changing the prior probability base on the local distribution and automate the process as well as finding the optimal confidence threshold. Implement the classifier in SVM to compare it with bayes classifier.