Machine Learning Approach for Answer Detection in Discussion Forums: An Application of Big Data Analytics

Table 1

All twenty features with brief description.

Code	Abbreviation	Description	Feature type	Subtype
F1	ThrdCntrodRplyWMDistance	Word mover distance of a reply from the thread centre	Lexical	Semantic
F2	ThrdCentrodRplyCosnSmlrty	Cosine similarity of a reply with the thread centre	Lexical	Pure lexical
F3	TtlRplyCosnSmlrtyWholCrps	Cosine similarity of a reply with the title based on corpus created from all threads	Lexical	Pure lexical
F4	QustionRplyCosnSmlrtyWholCrps	Cosine similarity of a reply with the initial post based on corpus created from all threads	Lexical	Pure lexical
F5	TtlRplyCosnSmlrty	Cosine similarity of a reply with the thread title	Lexical	Pure lexical
F6	QustionRplyCosnSmlrty	Cosine similarity of a reply with the thread initial post	Lexical	Pure lexical
F7	UnqWrds	Number of unique words in a reply	Lexical	Pure lexical
F8	IsRplyByCrtrOfInitlPost	Was the reply given by the creator of initial post?	Nonlexical	Structural
F9	NumRepliesByUsrCurrentThrd	Total number of replies given by the user in the current thread	Nonlexical	Structural
F10	NoThrdsUsrParticipated	Total number of threads the user has participated	Nonlexical	Structural
F11	ReWrdsOvrlpInitialPost	Number of overlapping words between the initial post and the reply post	Lexical	Pure lexical
F12	ReWrdsOvrlpThrdTitl	Number of overlapping words between the thread title and the reply post	Lexical	Pure lexical
F13	IsRplyContan5WHWrds	Does the reply contain 5WH words?	Nonlexical	Content based
F14	IsRplyMntionOthrUsrNames	Does the reply refer to any other forum user?	Nonlexical	Structural
F15	IsRplyHvHyperlnk	Does the reply have any Hyperlink?	Nonlexical	Content based
F16	WMDbtwnTitlRpl	Word mover distance between thread title and reply	Lexical	Semantic
F17	WMDbtwnQustionRpl	Word mover distance between initial post and reply	Lexical	Semantic
F18	TotlNoRpliesByUsrInAllThrds	Total number of replies given by the user in all threads	Nonlexical	Structural
F19	TotlNoIntialPstsByUser	Total number of initial posts created by the user	Nonlexical	Structural
F20	NoWrdsRply	Total number of words present in a reply	Lexical	Pure lexical