Mypesananlive: The development of facebook live order management system using natural language processing approach

• Main Author: Muhammad Danial Aiman Mohd Hanif
• Format: Academic Exercise
• Language: English; English
• Published: 2022
• Subjects: QA76.75-76.765 Computer software
• Online Access: https://eprints.ums.edu.my/id/eprint/33334/1/myPesananLIVE%2C%20THE%20DEVELOPMENT%20OF%20FACEBOOK%20LIVE%20ORDER%20MANAGEMENT%20SYSTEM%20USING%20NATURAL%20LANGUAGE%20PROCESSING%20APPROACH.24pages.pdf; https://eprints.ums.edu.my/id/eprint/33334/2/myPesananLIVE%2C%20THE%20DEVELOPMENT%20OF%20FACEBOOK%20LIVE%20ORDER%20MANAGEMENT%20SYSTEM%20USING%20NATURAL%20LANGUAGE%20PROCESSING%20APPROACH.pdf; https://eprints.ums.edu.my/id/eprint/33334/

A natural language processing (NLP) integrated Order Management System (OMS) for live streaming merchants are not widely developed, especially for small and medium businesses. The commonly used method for collecting orders from customers is screenshotting the Customer’s buying action comments (BACs) along with the snapshot of the live stream for the order context. This method will become much difficult when the number of orders reaches hundreds or thousands, requiring a significant number of human resources. Another problem is that orders through comments can be missed due to typos and incorrect format by customers when commenting, leading to orders not being detected by existing OMS systems. Not many existing OMS were developed with NLP integrated to identify BACs to automate the process of capturing orders. Therefore, a Facebook live order management system with natural language processing was proposed to resolve these issues. Text tokenisation with rule-based approach was utilised to capture orders from Facebook live stream comments. The objective of this project is: (i) To investigate the process of capturing order commands using a combination of rule-based algorithm and regular expression from live video comments and convert them into orders. (ii) To develop a Facebook Live order management system integrated with natural language processing techniques. (iii) To evaluate the system’s accuracy in detecting buying action comments from live streams. The expected outcome of this project is a fully functional Facebook live order management system.