Posted above is my word cloud made with my MALLET results. We had used MALLET previously in class and it was interesting to create a key word or category for a group of related words. Making them ourselves however was a different experience. I got to see what goes into making these topic models. I used 4 separate combinations when topic modeling. My first search was 50 topics/1000 iterations/ 20 words printed. Within this search I picked the 3 sentences that were able to be categorized the easiest. The topics for the three examples I chose were “Hallway”, “Communication”, and “Study/Office”. The second search I did was 25 topics/ 500 iterations/ 15 words printed. The three examples were “Case”, “Suspect” and “Evidence”. The third search I did was 20 topics/ 250 iterations/ 10 words printed. The four examples I chose were “Suspicious”, “Location”, “Discover/Trace” and “Attack/Violence”. During my search results I felt that it would be best to narrow my search requirements after every time. My reasoning behind this was that by narrowing my search queue, I would get more accurate results every time. I felt that the more words printed in a search results would make the topic harder to categorize because there is more words that you need to relate with each other. The models I got with narrower search results were easier to understand and easier to categorize. Overall, topic modeling using MALLET was a helpful tool to try and find main themes throughout all the Sherlock Holmes stories and I look forward to doing it again in class if given the opportunity.