Nuances of Data Pre-Processing and its Impact on Business
M. Sumathi, S.Ashika Parvin · 2021
Sentiment analysis (SA), is an open-ended field of research about computational remedy of classifying the human being's opinion about something. The data pre-processing plays a decisive position in SA which is obtained by accelerating the effectively classified instances through the selection of suitable preprocessing methods. When using both the Rule-based and Decision Tree algorithms separately, the missing of data happens that can cause the incorrect classifications of critiques and make the SA procedure not that effective. On the component of the above, this paper explains approximately the importance of pre- processing by introducing a novel approach of combining both the Rule-based and Decision Tree algorithms. The RBDT is a proposed approach which handles the lacking of information values effectively by making sure no word is missed during the classification process. The Experimental results classified the proposed technique to be more effective in achieving accuracy of 90% and time complexity of 0.079 seconds in predicting all the words in the dataset rather than using both the algorithms individually.