Load the Reuters Corpus provided in the NLTK library, which is a collection of news articles classified into different categories.
a. Use the Reuters Corpus to extract articles related to the "money-fx" and "coffee" categories.
b. Preprocess the articles by tokenizing them, removing stopwords, and performing stemming.
c. Create a document-term matrix for the preprocessed articles, where each row represents a document and each column represents a term.
d. Use Latent Semantic Analysis (LSA) to extract the top 10 latent topics from the document-term matrix.
e. Print the top 5 terms associated with each of the 10 latent topics, and interpret what each topic represents in the context of the "money-fx" and "coffee" categories.