Annotating and detecting topics in social media forum and modelling the annotation to derive directions-a case study

Journal of Big Data

Table 6 Precision (P), recall (R) and F1-scores of self-training classifiers with different training data sizes

Classifiers	Training size of the labelled data
	40%			50%			60%			70%
	P	R	F1-score	P	R	F1-score	P	R	F1-score	P	R	F1-score
NN (Base classifier)	0.682	0.691	0.684	0.688	0.697	0.686	0.704	0.708	0.706	0.729	0.714	0.711
NN (k = 0.85)	0.684	0.694	0.687	0.682	0.693	0.684	0.708	0.711	0.709	0.730	0.726	0.727
NN (k = 0.90)	0.720	0.723	0.721	0.731	0.730	0.730	0.741	0.740	0.740	0.749	0.746	0.747
NN (k = 0.95)	0.729	0.728	0.728	0.738	0.736	0.770	0.746	0.745	0.745	0.751	0.749	0.750