ziqilau / ndtcm

nDTCM for modeling social streams

nDTCM

Features

nDTCM is used to organize the data in social streams. We believe the meaningful semantic features discovered by nDTCM can also benefit the representation of social streams.
Using nDTCM as a building block can further analyze sophisticated structures/patterns in social streams.

Data

Each data point is represented as one line in plain format, splited by space, accordingly denoted as EPOCH #PARTICIPANTS PARTICIPANTS(split by commas) #DISTINCT_WORDS WORD:FREQUENCY 15 5 0,1,2,3,4, 18 14:1 43:1 45:1 31:1 40:1 49:1 6:2 7:2 33:1 34:2 22:1 2:1 8:1 32:1 26:1 27:2 53:1 23:1

Reference

[1] a dynamic nonparametric model for characterizing the topical communities in social streams. Ziqi Liu, Qinghua Zheng, Fei Wang, Zhenhua Tian, Bo Li.