Archiving System Optimization using Skip Gram based Neural Network as a Feature Selection

Automatic processing of massive unstructured data and extracting useful information are one of the challenges researched so far. Hence, the techniques of data mining are of great significance in this area and text categorization (TC) is one of those common researches. The key problem in TC is due to the huge amount of data. The lack of effective description of the text of the content and the implementation of appropriate features leading to a decline in categorization accuracy. Therefore, several methods appeared to select the features that help clustering and improving their performance as it appears in our research. We suggest a good way to select a feature, which support archiving system thus it supports the retrieval system.