Structured Support Vector Machine

In this paper, we propose the use of conditional random fields (CRFs) to address the challenge of image segmentation.As part of pre-processing the data, we perform oversegmention on the training images to represent them as a group of superpixels. By considering each superpixel as a vertex, we are able to utilize CRFs with improved computational efficiency compared to using individual pixels. We identified several unary features such as the superpixel color, histogram, and the histogram of oriented gradients. For the pairwise features, we considered the color difference, histogram difference, and texture similarity.We also discovered that the considering the location of the superpixels relative to the image had little effect in improving the performance of the model. By experimenting with the different combinations of unary and pairwise features for the model on the Weizmann Horse Dataset, we are able to develop a model that showed good accuracy.

Text mining and Text classification are the two prominent and challenging tasks in the field of Machine learning. Text mining refers to the process of deriving high quality and relevant information from text, while Text classification deals with the categorization of text documents into different classes. The real challenge in these areas is to address the problems like handling large text corpora, similarity of words in text documents, and association of text documents with a subset of class categories. The feature extraction and classification of such text documents require an efficient machine learning algorithm which performs automatic text classification.
This paper describes the classification of product review documents as a multi-label classification scenario and addresses the problem using Structured Support Vector Machine. The work also explains the flexibility and performance of the proposed approach for efficient text classification.