Hybrid Method for Text Graph Classification using 6DoS-CC Framework: A Case Study of Nigerian Hate Speech
Oluwatobi Balogun
ABSTRACT
In the use of the frequency of words sub-graph approach, the lack of instance feature representation constitutes a major obstacle. This experimental work makes use of the 6DoS-CC framework, a hybridization of the methods of the six degrees of separation for the validation of word graphs and the Collatz Conjecture for identifying vectorized frequency feature of hate-speech graphs based on the number of characters that form each word for the categorization of text. Results show a strong Pearson’s Correlation Coefficient value of 0.517462073 indicating a large strength of correlation figure based on the observed high disparity within odd hail stone numbers for Nigerian Hate speech.
