Text Analysis: A couple of ideas for your project:
--------------------------------------------------

1st version:
1. you need to figure out which criteria you want to consider to decide whether you like a text or not
2. define a measure (utility function) determining whether you like or not a text based on these criteria

2nd version:
Even if you end up with a list of criteria to evaluate texts, you can never know as much as when you read the texts for which you already know if you like them or not. So:
1. you will have to learn from the text for which you know if you like them or not (mor recurrent words for instance, length of the text, etc.)
2. from parsing these referent texts, you need to be able to isolate criteria that will be used for determining the degree of likeliness to like other texts
3. determine a utility function based on the isolated criteria

At this point you will be pretty well advanced, and you will have to design a strong testing strategy, a good way to report and analyze the results of your experiments.