Experiments With Text Mining

displaCy Named Entity Visualizer

Using the displaCy Named Entity Visualizer, I selected an excerpt from The Slang Dictionary, by John Camden Hotten. I noticed that the tool was able to accurately identify words under categories such as location and geopolitical entity, but seemed to classify a lot of terms under organization, many of which were incorrect. It seemed to make classifications based on the context of the situation or the way the word was presented, even though many words with very different associations are presented in the same way.

JSTOR Text Analyzer

Using the same text as above, this tool gave me reccomendations for other readings based on terms that I’m assuming the tool thinks are most prevalent. Based off of what I had taken from the text, I think the tool did a really nice job at coming up with 5 terms that summarize its main ideas. I was also able to change the weight of each term, I guess if I thought that one or more terms expressed really important ideas in the text.

Voyant

The Voyant tool found which words were the most common in the text and displayed them in the form of a word cloud. This could be great for seeing commonalities between different texts. There was also a section titled correlations. This showed terms that were in some way connected to each other. I’m not sure if this was based on definition or if the tool was able to recognize that when one term was used, it shortly after made reference to the other term, but regardless shows how certain pieces of information can be linked to another