Sentiment analysis and the complex natural language

Complex Adaptive Systems Modeling

Table 1 Comparison of the techniques and approaches included in the study

SNo.	Naïve Bayes	k-Nearest neighbor	Centroid
1	Yes	Yes	Yes
2	Yes	Yes	No
3	Yes	No	No
4	Word probability	Value of k	Centroid vector
5	Probability weights	Distance similarity	Vector distance
6	Simple and fast	Handle co-related features	Classify on vector distance
7	Assume feature independence	Sensitive to irrelevant features	Sensitive to noise
8	Yes	Too expensive	No

SNo.	Support vector machine	Lexicon (dictionary) based	Statistical (corpus) based
1	Yes	No	No
2	No	NA	NA
3	No	No	Yes
4	Kernel function	Word polarity	Feature matrix
5	Hyperplane	Word polarity	Word distance
6	Classify on hyperplane	Can identify new lexicons	Handle online data
7	Require more resources	Struggle with domain context	Conceptual document size
8	No	Yes	Yes

1 require training, 2 use training data to classify, 3 probabilistic approach, 4 driving factor, 5 similarity metric, 6 strength, 7 weakness, 8 support for streaming data