On: Show HN: An easy-to-use Text Analysis API – NLP an...

[cliveowen]

"There was a time when men could roam free on earth, free from concrete and tarmac. Now it's all gone to shit."

Classification: arts, culture and entertainment - architecture .(WTF?)

Polarity: positive. (Nope)

Polarity confidence: 0.9994709276706056. (Well...)

Looks pretty rough to me.

[guptaneil → cliveowen]

Why does that classification elicit a WTF? That seems like a reasonable classification, given how little context the algorithm has about the snippet. It's entirely plausible for that quote to be from a book about how "concrete and tarmac" have impacted modern architecture. There's not really any other hints about what it could be about.

There's no excuse for the polarity though. "Gone to shit" should be a pretty good indicator about the sentiment.

[parsabg → cliveowen]

thanks for the feedback. considering the fact that it's still a v1, it surely can be rough in some areas. anyway, here are some thoughts:

- classification is trained on longer texts (mainly news articles) so it won't perform well on shorter texts

- polarity: yeah that's bad - the sentiment is slightly ambiguous so maybe a lower confidence would've justified things

[jonahx → cliveowen]

Isn't it supposed to analyze something of article length, or at least paragraph length? Seems unfair to slam it for messing up a single sentence.