1) Prepare dataset for good product vs bad product query if required. ( One can do it quickly using [login to view URL] )
2) Examples of good product queries
- "nail polish to go with navy dress"
- "computer science book"
- "dog seat cover"
- "dog seat cover for ford f150"
- "iphone 7 yellow case"
- "multi purpose hammer"
- "hammer with rubber head"
- "Samsung phone's best cover"
- "General purpose hammer"
- "non toxic baby soap"
3) Examples of bad product queries:
- "battery life of iphone 7"
- "nail polish may kill you"
- "social work"
- "table dimension"
- "rock music"
- "computer science"
- "hammer broke iphone 7"
- "nokia phone stronger than hammer"
- "iphone 7 damaged by new case"
- "Samsung phone looks ugly with cover"
Criteria for success: Will randomly query suggestions from [login to view URL] and check how the algorithm is working.
Note 1: Attached files, I got from [login to view URL], by using some seed words like: "battery life", "nail polish", "rock music", "social work", "dog seat cover", "iphone 7" and "hammer" . I have a paid account for answerthepublic.com. Do let me know if you want more data for seed queries.
Note 2: I tried to remove few bad queries by using some stop-words like "you", "my", "won't", etc., And also using bit of nltk based POS tagging and considering only queries ending with nouns. But had minimal success. Looking for a superior solution with high precision [ Its ok if you loose few good queries, but classifying bad query as good query has to be very minimal ]
Note 3: Feel free to use any of the nlp libraries like spacy, nltk etc.,
Additional task: Convert the query to product if possible
"social work" to "social work book"
"rock music" to "rock music cd"
"computer science" to "computer science book"
13 фрилансеров(-а) в среднем готовы выполнить эту работу за ₹25160
Hi I read your project description. I had worked with nltk token analyzer to map tokens for sentence generation and semantic analysis. I would like to discuss the project, feel free to contact me.
My experience in working with advanced ML models and improved algorithms guarantees trying a wide range of different approaches in order to achieve the best results for the project.
I can make your projects within minutes with efficiently and effectively. I have good command on NLP projects. I am Graduate student and have done like these kinds projects as sentimental analysis etc.