INTERACTIVE SEGMENT EXTRACTION IN COMPUTER-HUMAN INTERACTIVE LEARNING

摘要

Extremely large collection of data can make it difficult to search and / or analyze. Relevance may be dramatically improved by automatically categorizing queries and web pages in useful categories and using these classification scores as relevance features. A thorough approach may require building multiple classifiers corresponding to various types of information, activities, and products. The generation of classifiers and schemers is provided on large data sets. Practicing classifiers and schemers for millions of items may expose data-specific values by adding available meta-data. Some aspects include active labeling search, autotemplation and coldstart, scaling with number of items and number of classifiers, active feature, and segmentation and schemaization.