Labelling Training Samples Using Crowdsourcing Annotation for Recommendation

Table 1

Symbols.

No.	Symbols	Presentation
1	K	A keyphrase
2	T	A training sample
3		An independent keyphrase
4		The attribute meaningfulness
5	#NKI	The number of occurs
6		A sub-keyphrase
7		The attribute uncertainty
8	#NKS	The number of occurs
9	#TN	The total number of keyphrases in the corpus
10		The attribute uselessness
11	H (k)	The information entropy of k