Publication: Trashy : A smart chatbot for sustainable trash management
2
0
Issued Date
2021
Resource Type
Language
eng
File Type
application/pdf
Access Rights
open access
Rights
ผลงานนี้เผยแพร่ภายใต้ สัญญาอนุญาตครีเอทีฟคอมมอนส์แบบ แสดงที่มา-ไม่ใช้เพื่อการค้า-ไม่ดัดแปลง 4.0 (CC BY-NC-ND 4.0)
Rights Holder(s)
มหาวิทยาลัยศรีนครินทรวิโรฒ
Suggested Citation
Pattamas Wiratchanee, Wutthipong Ranmeechai, Suwikan Wongwean (2021). Trashy : A smart chatbot for sustainable trash management. Retrieved from: https://hdl.handle.net/20.500.14740/10037
Advisor(s)
Organization
Abstract
Recently, waste management is one of the biggest problems encountered by many countries. Improper waste management leads to many ensuing problems for society such as health issues due to germs and global warming since many people destroy waste by burning. One possible solution to alleviate this problem is to properly classify the type of garbage and handle each type efficiently. For example, some organic waste could be fermented to be used as fertilizer or recycle waste for sale to reuse for maximum benefit. However, waste classification in many countries is neglected due to many factors such as incompetent authority or lack of waste management awareness leading to inefficient waste management system. In this project, we propose "Trashy”, a smart chatbot for sustainable trash management. Trashy is an intelligent software chatbot that gives suggestions to users on how to responsibly manage waste. Trashy employs a deep learning model called resnet50 to analyze a picture of garbage and classify them into 7 types: glass, plastic, metal, paper, general waste, food waste and hazardous waste. Once the garbage has been classified, Trashy advises users on how to properly take care of the garbage such as identifying the proper type of bin to dispose of the trash or notifying suitable places and time to sell recycled waste. Preliminary results show that our trash classification model yields 98% classification accuracy where each image takes approximately 4 seconds to process. In addition, Trashy uses a deep learning model called LSTM to predict trash price with root mean square error (RMSE) equal to 0.07 baht/kg based on the past waste price.