<To-do list>
1. Contens based on recommend system design
1.1) 사진기반 분류 가능여부 판단 (v)
1.2) 데이터셋 만들기. ( )
2. Should I build Ontology sets? not yet
3. type of recommend system - User, Contents, Collaborate
4. How can I visualizing this Model?
5. 내가 쓴 글들이나 목록을 통해서 책을 추천해 줄수도 있겠다.
<모듈화 계획>
1. Extracting 3 keywords from photo using Goolgle Vision Api
2. Classification Model using Doc2vec model
0) Doc2vec trainnig - 50000 reviews each category
1) finding sim_word_vector
2) finding sim_doc by sim_word_vector
3) predict what category is related with the photo.
3. Building recommend system module
1) contents based recommend (which big category you would like) - macro
2) collaborate based recommend (which product line in the category you would like) - micro