Reference no: EM132496418 , Length: word count:1600

Assignment - Data Science Questions

These questions are to be answered in addition to your main data science project. All questions are to be answered using your dataset that you extracted for your own project.

References are not required for this assessment.

Q1. Conduct a sentiment trajectory analysis on (at least) five authors in your dataset. How do the average trajectories differ?

Q2. Use q-gram-based similarity calculations and analyse (for at least two authors) whether the internal similarity (i.e. the similarity among texts of a single authors) is higher than the similarity across authors (i.e. the similarity between texts written by different authors).

Q3. Extract the overall sentiment of each text in your corpus (i.e. a single score for each text). Now build and run a supervised classification model and report detailed performance metrics (i.e. accuracy, precision, recall, F1 score, and area-under-the-curve).

Q4. Choose five texts - each of a different author. For each text find another text in your whole corpus that is most similar in terms of (i) unigram usage, (ii) word length, and (iii) the occurrence of stop words?

Note - Each question must be answered within 400 words.