For RWET final I made a recipe book of life advice.
1.Scrape answers from Quora
I give it the URL of a particular question, and it will return me the scraped answers for the question, and save them as a text file.
2. Extract all the nouns (of certain relative dependency) and clean the data
3. Create the framework for generating fake recipe
part 1: Ingredients — (randomly select 15 words from the ingredients pool (the list of nouns from the previous step)
Part2: staring step — Creating a list of starting word for running Markov chains (prepare, cut, chop etc.)
Part3: middle steps — Creating a list of starting words for running Markov chains
Part3: last step — Creating a list of starting words for running Markov chains (Finally xxxx)
4. Run Markov Chain
I tried different source texts including scraping from recipe websites, various cookbook. The one I ended up using is an Italian cookbook.
4. Replace original ingredients with new ingredients from the list of nouns we got from quora scrape
Still looking for a way to better identify the ingredients.
Some of the results: