Signal and Data Processing

Download citation:
BibTeX | RIS | EndNote | Medlars | ProCite | Reference Manager | RefWorks
Send citation to:

---

Today, with the increasing use of social networks, the amount of data generated is increasing. On the other hand, many businesses are active in different social networks; For this reason, recognizing the needs of users for marketers in social networks is one of the requirements for the development of internet businesses and e-commerce. Therefore, it is important to automatically detect request messages and filter them in a way in Persian texts. The present research was carried out to improve the recognition of request messages in the set of messages sent in messengers. Nowadays, social networks are easily accessible; Therefore, messages in social networks are different from literary texts. Messages in social networks have additional and popular data. On the other hand, the words contain many spelling mistakes; Therefore, dealing with these messages is considered a challenge. In this research, pre-processing and removal of additional data have been investigated. To deal with other challenges raised, the proposed method has also dealt with spelling mistakes while maintaining the value of words. After extracting the appropriate features, a hybrid model based on deep neural networks was designed for the process of recognizing and classifying request messages. In the evaluation phase, comprehensive tests were implemented to analyze the performance of the proposed model. According to the obtained results, the precision, recall, and f-score of the proposed method is almost 90% and compared to the previous methods presented, it was improved by 5% on average.

Keywords: E-commerce, request detection, social networks, messenger, deep learning based method

Full-Text [PDF 929 kb]   (25 Downloads)    

Add your comments about this article : Your username or Email:

---

---