Vectorizing text messages and training a multinomial Naive Bayes and AdaBoost classifiers to identify text messages as either "spam" or "ham".

Lemmatizing text and using BeautifulSoup to visualize the most common words used in positive and negative reviews. Training a Logistic Regression with a threshold of 0.5 to identify positive and negative reviews. Calculating and identifying the misclassified reviews.

Using PCA and LSA to reduce dimensions of tokenized text vectors. PCA is used to sort book title keywords into a two-dimensional vector, which shows the keywords having two main axis - o containing social/historical keywords; other - scientific and data-driven.

Modifying articles with randomly selected possible words based on the 2nd order Markov's assumption model. The text is imported from html, converted into tokens, and then - into a trigram dictionary, where every two surrounding words have the possible encountered middle words and their counts (later converted to probabilities) to appear in text.

• Pandas
• Numpy
• NLTK
• matplotlib
• sklearn
• wordcloud
• bs4 (BeautifulSoup)