A termsetA termset - also known as itemset [16] or word combina- tion feature [28] - is assumed to occur in a given document if all members are present, regardless of their order and position. Selecting a discriminative set of n-termsets is a highly crucial, but very challenging, task since all groups of n terms can be candidate n-termsets. In order... – also known as itemsetA termset - also known as itemset [16] or word combina- tion feature [28] - is assumed to occur in a given document if all members are present, regardless of their order and position. Selecting a discriminative set of n-termsets is a highly crucial, but very challenging, task since all groups of n terms can be candidate n-termsets. In order... [16] or word combina- tion feature [28] – is assumed to occur in a given document if all members are present, regardless of their order and position. Selecting a discriminative set of n-termsets is a highly crucial, but very challenging, task since all groups of n terms can be candidate n-termsets. In order to sim- plify this process, one can include only frequent terms in generating termsets [29]. Defining termsets as sets of dis- criminative terms has also been studied [15]. Alternatively, using various feature selection methods such as χ2, mutual information, odds ratio and information gain are evaluated for this purpose [16]. The use of both frequency of termsets and their distribution across different classes has also been studied [30].

 

Badawi, D., & Altınçay, H. (2017). TermsetA termset - also known as itemset [16] or word combina- tion feature [28] - is assumed to occur in a given document if all members are present, regardless of their order and position. Selecting a discriminative set of n-termsets is a highly crucial, but very challenging, task since all groups of n terms can be candidate n-termsets. In order... weighting by adapting term weighting schemes to utilize cardinality statistics for binary text categorization. Applied Intelligence, 47(2), 456–472. http://doi.org/10.1007/s10489-017-0911-6