Multi-class Weather Forecasting from Twitter Using Machine Learning Aprroaches

Now in the era of big data, many are applying information methods accurately especially by social media. The aims of this study to classify the weather based on Twitter automatically using text mining by using Support Vector Machine (SVM), Multinomial Naive Bayes (MNB), and Logistic Regression (LR) method. The experimental results show that SVM substantially outperforms various other machine learning algorithms for the task of text classification with an accuracy value of 93%. This result proves that SVM is very suitable for text categorization. We use clustering technique to read the pattern in customers’ opinion about the restaurant based on some measurement variables.

International Conference on Computer Science and Computational Intelligence 2020

Kartika Purwandari, Join W C Sigalingging, Tjeng Wawan Cenggoro, and Bens Pardamean

Read Full Paper