| Recognition of joy and sorrow in Persian texts by using Single-layer Neural Network (Perceptron) |
| Paper ID : 1078-ICIL |
| Authors: |
|
Parvaneh Khosravizadeh *1, Mohammad Rajabpour2 1Sharif University of Technology, Azadi Street 2Department of Computational Linguistics, Sharif University of Technology |
| Abstract: |
| The purpose of this research is data mining, recognition, and analysis of words that express happiness and sadness in linguistic data by using computational methods. The analyzed samples in this study were texts that represented the feelings of joy and sadness in cyberspace. At the first phase of this study various texts including words representing greetings, condolences and sympathy were retrieved in the official and informal language context from the Internet. Consequently, the results were placed in two categories namely ones represented joy and those represented sorrow. The present research study was carried out using training and test methods by taking advantage of single-layer neural network (perceptron). The results showed that just by using single-layer neural network unseen data can be classified to happy and sad text categories with a high range of accuracy. The results of this research can be used in studies related to information retrieval, data mining, and Persian language processing. |
| Keywords: |
| data mining; sentiment analysis; computational linguistics; neural network; perceptron |
| Status : Paper Accepted |