2018 SNU-FIRA AI Agent Course, Final Capstone Project (18.11.06-12.14)
This project is supported by Seoul National University Big Data Institute.
๐ Our team won 1st place in the final project presentation.
Building Product Matching Models for an Ecommerce Platform(KOR) using Deep Learning Methods.
โโโโCLOSED DATASETSโโโโ ํ๋ก์ ํธ ํ๋ ฅ๊ธฐ๊ด๊ณผ์ ๋น๋ฐ ์ ์ง ํ์ฝ์ผ๋ก ๋ฐ์ดํฐ๋ฅผ ๊ณต๊ฐํ ์ ์์ต๋๋ค.
Initial requirements are as follows.
python 3.6.5
gensim 3.6.0
keras 2.2.4
tensorflow 1.12.0
numpy 1.15.4
pandas 0.23.4
*---main.py---*
โ | ํ๋ง๋ผ์ผ ๊ณ ์๋ถํฌ๋ฆผ (์ธํ
์๋ธ) 150ml ===> ์ธํ
์๋ธ ๋ชจ์ด์ค์ฒ๋ผ์ด์ง ํฌ๋ฆผ@150ml
โ | [HANYUL]์ด๋ฆฐ์ฅ ์๋ถ ์ง์ ํฌ๋ฆผ 50ml ===> ์ด๋ฆฐ์ฅ ์๋ถ ์ง์ ํฌ๋ฆผ@50ml
โ | ๋ง๋ชฝ๋ ๋ชจ์ด์ค์ฒ ์ธ๋ผ๋ง์ด๋ ์ธํ
์ค ํฌ๋ฆผ ===> ๋ชจ์ด์ค์ฒ ์ธ๋ผ๋ง์ด๋ ์ธํ
์ค ํฌ๋ฆผ@50ml
โ | ํ ๋๋ชจ๋ฆฌ ๋ ์ด์ด ๊ทธ๋ฆฐ ํฐ ์๋ถ ํฌ๋ฆผ 60ml ===> ๋ ์ด์ด ๊ทธ๋ฆฐํฐ ์๋ถ ํฌ๋ฆผ@60ml
โ | ์์ด์คํ ๋๋ง ๋ฆฌํ์ด ์์นดํฌ๋ฆผ ===> ๋๋ง ๋ฆฌํ์ด ์์นดํฌ๋ฆผ@50ml
โ | ํค๋ผ D_ํค๋ผ ๋ก์ง ์ฌํด ํฌ๋ฆผ 50ml ===> ๋ก์ง ์ฌํด ํฌ๋ฆผ@50ml
โ | ๋ง๋ชฝ๋ ๋ชจ์ด์ค์ฒ ์ธ๋ผ๋ง์ด๋ ์ธํ
์ค ํฌ๋ฆผ 50ml -์ ๋ฌผํฌ์ฅ1 ===> ๋ชจ์ด์ค์ฒ ์ธ๋ผ๋ง์ด๋ ์ธํ
์ค ํฌ๋ฆผ@50ml
โ | (27%ํ ์ธ)์คํ์ ํ๋ ฅํฌ๋ฆผ 75ml ===> ํ๋ ฅ ํฌ๋ฆผ@75ml
โ | [์นด๋ 5% ํ ์ธ][CJmall]์์ด์คํ [๊ตฌ๋งค๊ธ์ก์ฆ์ ์ ์ธ]์์ด์คํ ๋ชจ์ด์คํธ์ ํฌ๋ฆผ ์คํจ ํ์ด๋๋ ์ด์
50ml ===> ๋ชจ์ด์คํธ์ ํฌ๋ฆผ ์คํจ ํ์ด๋๋ ์ด์
@50ml
โ | ํ๋ง๋ผ์ผ ์ ํ-50ml ํ๋ง๋ผ์ผ์ธํ
์๋ธ์๋ถํฌ๋ฆผ/ํ๋ง๋ผ์ผ์ธํ
์ ===> ๋๋ฆฌ์ฑ ์คํจ ํฌ๋ฆผ@50ml | โช์ ๋ต: ์ธํ
์๋ธ ๋ชจ์ด์ค์ฒ๋ผ์ด์ง ํฌ๋ฆผ ...
- Joulin, Armand, et al. "Fasttext. zip: Compressing text classification models." arXiv preprint arXiv:1612.03651 (2016).
- Shah, Kashif, Selcuk Kopru, and Jean David Ruvini. "Neural Network based Extreme Classification and Similarity Models for Product Matching." Proceedings of the 2018 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, Volume 3 (Industry Papers). Vol. 3. 2018.
- How to predict Quora Question Pairs using Siamese Manhattan LSTM