Text Classification Using Convolutional Neural Networks

Abstract

Most of the information (more than 80%) is stored as text, and text mining is a very important process as it is an initial step in the process of text classification, and this is especially the case in the Arabic language. The Aim of The Study is to classify Arabic texts according to specific categories using advanced performance indicators We used Data Templates as a platform for managing and organizing Apache Spark to solve big data challenges. Apache Spark offers several integrated language APIs. nlp lib was used for text processing. The data is preprocessed through several steps, namely separating the words into one text on the basis of the space between words, cleaning the text of unwanted words, restoring the words to their roots, as well as the feature selection process is a critical step. in text classification. It is a preprocessing technology. In this paper, one way to determine which TF attributes are used how often each feature appears in the document is that they consider the first level of the feature selection process. Then we use TF-IDF to determine the significance of the feature in the document, and this is the last step in the preprocessing Outcomes Text classification . Results were evaluated using advanced performance indicators such as accuracy, Precision and recall. A high accuracy of 96.94% was achieved. The main objective of this paper is to classify basic texts quickly and accurately, according to the results as long as the feature size is suitable, the most advanced technology is superior to other pass rate methods due to the reasonable reliability and perfect pruning level.