Article Spinning Software

I have a page spider that goes out and returns me articles that it finds on other web pages written in python.

Right now I have two different algos to spin the articles, one uses a translation service from Google, and the other uses NLTK.

1) the first algo tolkenizes the sentences and re-orders them, next I replace adjectives and adverbs with synonyms using NLTK.

2) this algorithm translates an article to another language and then translates it back to english.


What is the best way to spin articles using Python?

Additionally if I were to publish like 100,000 unique articles per day how do you think this would do for page views?

Best,

dont have any idea about python

If you are looking for bulk spinning you can either use your own synonym database or grab Wordnet's http://wordnet.princeton.edu/. If you want a little better quality you can use The Best Spinner API as they allow up to 1.25 million words processed per day (albeit in 5000 word chunks). If you want to go crazy with daily volume though you should just make a markov script and feed all of the articles you are scraping into the corpus index.

I have been using wordnet for NLTK, it works great but I am not sure how much I need to obsfuscate the articles in order for Google to think they are original.

My spider sends back plaintext articles, and I am not sure if I should paraphrase the article, and post it, or paraphrase it and spin it multiple times, and posting it multiple times?

I have more than enough raw content, so I think that how I paraphrase it is the key.

I personally don't think google will be able to see duplicate content if I re-order the sentences and then switch out some synonyms.

What has worked best in your experience for paraphrasing?