Page 69 - IJOCTA-15-1
P. 69
BSO: Binary Sailfish Optimization for feature selection in sentiment analysis
(a) F-score values
(b) Execution time
Figure 3. F-score and execution time comparisons of ML algorithms according to k value
Figure 4. The results of preprocessing combination
4.3. Binary Sailfish Optimization (BSO) also implement binary versions of the Harmony
for SA Search (HS) 89 and Bat Algorithms (BA), 90 both
of which have been successfully applied to various
In this section, we employ the BSO algorithm as optimization problems. 91,92 Besides these, we ex-
a feature selector for the first time in the con- tend our performance evaluation by including the
text SA. To evaluate the performance of BSO, we Atom Search Optimization (ASO) algorithm, 93 a
63
