sentiment.analysis		package:none		R Documentation

Gives a sentiment score to a text that is the sum of the score of all scored words/expressions in the text.

Description:
	
	sentiment.analysis receives a text and a table of words
	and their scores (if none is provided, the function will
	search for the "WordScores.txt" file provided with this
	function). Each word and expression of up to 4 words from
	the text is searched in the table data base and the scores
	of all words/expressions is added to get the text score.

Usage:
	
	sentiment.analysis(your.text, word.list.and.scores)
	
	## Default:
	sentiment.analysis(your.text, word.list.and.scores="WordScores.txt")

Arguments:
	
	your.text:	character. A text given by the user in
			UTF-8 encoding, in english language,
			usually a .txt file, but internet sites
			can also be given here.

	word.list.and.scores:	table. A table with the words/
				expressions in the fist column,
				and their score in the second
				column. The separator must be
				a tab ("\t") and their should
				be no header in the file.

Value:

	The function returns the value of the text score (sum of all
	word/expression scores) and two graphics: one of the relative
	positive and negative words (in the universe of all categorized
	words) and one of the relative positive, negative and neutral
	words (in the universe of all words from the text, were un-
	categorized words are considered neutral).

Warning:
	
	The function is not really fast, the bigger the text and
	you word list and score the more it will take. With the
	default word list even small texts may take a while to be
	processed. Remember that R demands quite a bit from your
	RAM memory, so it mey be a good ideia to make a camomile
	tea while you wait for the function to run, mainly if you
	are using large data.
	
	Also, in the case of low RAM memory and large texts there
	may appear a few warnings, but the program usually works
	out just fine. Just don't forget the camomile tea.

Author:

	Júlia Beck Raíces
	nºUSP: 6802291
	
	julia.raices@gmail.com
	juliar@riseup.net
	fingerprint: BF75 AF9A 1232 DFF6 0189 5D72 7877 3E81 1433 5F11

Thanks:

	Special thanks to Viviane Santos who helped me with the ideia
	of the function and with the references and to Chalom, who
	always helps me with the constant despair of computer programming.

References:

	- The words list and scores was obtainned (and slightly modified) from:
	" Lars Kai Hansen, Adam Arvidsson, Finn Årup Nielsen, Elanor Colleoni,
	Michael Etter, "Good Friends, Bad News - Affect and Virality in
	Twitter", The 2011 International Workshop on Social Computing,
	Network, and Services (SocialComNet 2011). "

	- The "stoya.txt" test archive is the text "Sigh" from Stoya.
	Obtained from her blog at: http://graphicdescriptions.com/11-sigh

See Also:

	- Bo Pang and Lillian Lee "Opinion Mining and Sentiment Analysis",
	Foundations and Trands on Information Retrieval, Vol 2 (2008).

	- SentiWordNet ( http://sentiwordnet.isti.cnr.it/ )
	
	- Stanford's Sentiment Analysis website
	( http://nlp.stanford.edu/sentiment/ )

Examples:

	# Download both the "WordScores.txt" file and the "stoya.txt"
	## file at http://tinyurl.com/q353stg
	
	sentiment.analysis("stoya.txt", "WordScores.txt")
	# gives the text score and the graphs

	
	# Download both the files "WordScores.txt" and "test.txt"
	##at http://tinyurl.com/q353stg

	sentiment analysis("test.txt")
	# gives the score and graphics for another text (in the case
	## a made-up text of scored words). Notice that when
	## word.list.and.scores is not given the function automatticaly
	## uses the "WordScores.txt" file.