Just using lots of synonyms can obfuscate the text, and detract from its richness.

The richness of writing can be analysed (and operationalised) as its:

• high quality.
• style,
• harmony,
• total size of corpus (vocab),
• vividness (use of colour, sound, adjectives ...),
• yielding pleasure (evoking emotions),
• use of vigourous images (and phrases).

Thank you Mr Ian Kennedy for the answer. In what context you mean by "High quality"?

The novelty (originality) of ideas in the article

High quality is less easy to operationalise.

* It keeps on asking "what does the reader want and need?". It is reader friendly.

* It is precise and factually correct.

* The reader learns something or experiences something from reading it.

* It gives closure (What it asks or hints at in the beginning is answered by the end.)

* It is polished, with the most appropriate synonym consistently being used according to the context.

* It is well structured (Intro, body, conclusion etc.) and well developed.

* It is parsimonious.

* It is fresh as a daisy in Spring.

Very nice and to the point answer by Prof Kennedy.

Thank you verymuch Prof. kennedy.

Actually, I am working on a project in which I am trying to assess the difficulty level of an article. I am trying to use Machine Learning. Though, there were many research papers on this issue, I found myself missing some points in assessing the article. So, I thought to do it from the scratch and trying to find which features to consider to get best results.

If you/any others who is reading this, it will be great help to me, if you people can share some related articles.

I think that it would be difficult to analyze this quantitatively.

I think that what you call "richness" is the result of identifying a worthwhile research problem, collecting and analyzing data that effectively addresses the problem, and then providing an extensive discussion of how the findings fir in with other research...and recommendations resulting from the findings.

These are not things that can be determined from word or adjective counts, although you could probably create a rubric to quantify the qualitative scrutiny of the studies...with multiple coders requiring good inter-coder reliability.

Thank you Prof. Marek.

Of course you are right it is not sufficient to just count the Parts of Speech, probably looking into tenses i mean syntactical, semantics features etc.. would help.

I am thinking what exactly should I include in my feature set to estimate the richness of the article, till now I came up with POS, tenses, different phrases. Any other suggestions are indeed very helpful.