If I Loved Natural Language Processing Less, I Might Be Able to Talk About It More | Julia Silge

[utterances-bot]

If I Loved Natural Language Processing Less, I Might Be Able to Talk About It More | Julia Silge

In my last post, I did some natural language processing and sentiment analysis for Jane Austen’s most well-known novel, Pride and Prejudice. It was just so much fun that I wanted to extend some of that work and compare across her body of writing.

https://juliasilge.com/blog/if-i-loved-nlp-less/

[ozengnr]

Hello Julia,

Thanks so much for this -- I really enjoyed your work and now I'm trying to do a similar analysis. My dataset consists of 2 columns: a text feedback and a date that the feedback was given. I was able to plot the feedback line by line and aggregate by sentiment. However, I'd love the X axis to be Date, rather than the line number so I can look at it monthly/quarterly/yearly.

Is there a way to group these lines by day?

I appreciate it!

Oz

[juliasilge]

Yep, absolutely! Take a look at this section of our book, and group_by() your date variable instead of index or line number.

[mohamedelhilaltek]

hello julia I am currently participating in a text binary classification competition and have observed a 70% similarity in text between the training and testing datasets. How can I leverage this similarity to my advantage?

[juliasilge]

@mohamedelhilaltek I recommend that you take a look at Supervised Machine Learning for Text Analysis in R for best practices in binary classification with text.