RATINGS ARE OVERRATED
‘Service is great, but the desserts are bad”. Overall rating 4.5/5.
Many times we have gone into a restaurant alone or in a group, came out happy and still rated it a 3 or a 4 on social media. Does any one check why 2 precious points were deducted? No one will read your review unless the overall rating dips below 3.5. Is it an healthy practice? Of course not.
Frankly, no one is to blame except the rating scales and NPS…
Added by Manas Ranjan Kar on December 28, 2015 at 7:57am — No Comments
Natural Language Processing (NLP) is a messy and difficult affair to handle. Preprocessing, machine learning, relationships, entities, ontologies and what not.
Word embeddings/representations – ever since they came in with great work of Mikolov et al, they have been revolutionary to say the least. The concept itself is very intuitive and motivates deeper understanding fora wide range of applications. The…Continue
Added by Manas Ranjan Kar on December 28, 2015 at 7:53am — No Comments
Sentiment analysis is hard. Most of the systems on the market will clock anywhere around 55-65% for unseen data, even though they might be 85%+ accurate in their cross-validations.
A couple of reasons why creating a generic sentiment analyser is tough;
- There is too much variation in texts across domains, leading to different meanings
- Identifying sarcasm and combination of phrases like, 'not bad' is not equal to 'not' AND 'bad'