Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

How are you calculating percentages?


We used an open source sentiment analysis API.


This could perhaps be improved. It seems to employ a very primitive negative/positive, large/small dichotomy. I presidct that simply putting the words huge and new will give this comment a positive sentiment....


When comments like 'I'm pretty stoked about it actually.' yield a -50% sentiment, something is broken.


Which one?


text-processing.com[1] has a nice api, the bonus was that they restrict API calls on a per-IP basis. So client-side extensions work well here. :)

(meta: Let's see how much a smiley affects the positivity...)

[1] http://text-processing.com/docs/sentiment.html


Annoyingly their premium link 404s, here's the pricing link for anyone interested: http://www.mashape.com/japerk/text-processing/Pricing

Free plan is 45k queries, $35/m for 450k and 1,500,000 for $75/m!




Consider applying for YC's Winter 2027 batch! Applications are open till November 2.

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: