An Efficient Method for Noisy Annotation Data Modeling
Journal Title: IOSR Journals (IOSR Journal of Computer Engineering) - Year 2014, Vol 16, Issue 5
Abstract
Abstract : Probabilistic topic models are used for analyzing and extracting content-related annotations from noisy annotated discrete data like WebPages on WWW and these WebPages are stored using social bookmarking services with the help of social bookmarking services, reason behind this process most of time users can attach annotations freely, some annotations do not describe the semantics of the content, therefore they are noisy, simply they are not content related. The extraction of content-related annotations can be used as a prepossessing step in machine learning. Prepossessing step in machine learning is like text classification and image recognition, and can improve information retrieval performance. The proposed model is a generative model for content and annotations, where annotations are assumed to be originated either from topics that generated the content or from a general distribution unrelated to the content. We demonstrate the effectiveness of the proposed method with the help of synthetic data and real social annotation data for text and images
Authors and Affiliations
Sushama Shinde , Shyam Gupta
Keywords
Related Articles
- EP ID EP152688
- DOI 10.9790/0661-16512024
- Views 187
- Downloads 0
‘A Review Study on Future Applicability of Snake Robots in India’
Abstract: In this study we to aim to present an overview of features of snake robots and their application across various fields. Snakes is blessed with a unique feature of moving over or climbing all most all kind of te...
Validation Experimentations of Local Alignment Parameters for Comparing DNA and Protein Sequences
Abstract : A basic issue in aligning DNA and protein sequences is to find similar characters between two or more sequences in order to detect relations between newly defined sequences and well-known sequences store...
A Novel Irreversible Transformation Scheme for Biometric Template Protection
Abstract: Modern biometric technologies claim to provide alternative solution to traditional authentication processes. Even though there are various advantages of biometric process, it is vulnerable to attacks which can...
Energy Efficient Query Optimization in WSN usingThreeLevel Modelling
Abstract:With advancement in microprocessor technology, wireless sensor networks is playing vital role in different types of applications such as health care monitoring, fire detection etc. In these applications, large a...
An Entropy-based Feature in Epileptic Seizure Prediction Algorithm
Abstract : Epilepsy prediction is a vital demand for people suffering from epileptic onset. Prediction of seizure onsets could be very useful for drug-resistant epileptic patients. We propose an epileptic seizure predict...