2 ms·
At what scale? Until you've reached the point where you're receiving hundreds of bits of user feedback per day, you can just read the comments directly. Or rand
by PragmaticPulp 6y ago
At what scale? Until you've reached the point where you're receiving hundreds of bits of user feedback per day, you can just read the comments directly. Or random sample them.
Don't bother with sentiment analysis. Give users a single axis to directly indicate their satisfaction. Rate an app with a number of stars, or ask Net Promoter Score questions. Keep it simple.
- ChefboyOG 6y agoThis. As with many things in applied ML (although, importantly, not all), the question of "is x useful?" comes down to the scale at which you're operating.
- sanmak 6y agoThanks for the suggestion. Keeping in simple is the crux. In my case, there is a text box where user have to write, and on top of that we are doing the analysis.