Memuat
Memuat
Turn Indonesian customer reviews into positive or negative sentiment.
1 views
This project classifies 2,400 fictional Indonesian product reviews as positive or negative. Text is turned into word attributes with TF-IDF weighting, Naive Bayes and SVM are compared, and the most influential words are shown for the marketing team.

Read CSV loads ulasan_produk.csv (2,400 reviews). The ulasan column is read as text, sentimen as a binominal label and id_ulasan as the id.
Remap Binominals sets "positif" as the positive class and "negatif" as the negative class.
Generate Attributes creates one 0/1 column per keyword (e.g. w_bagus, w_kecewa) with the formula if(contains(lower(ulasan),"bagus"),1,0). Phrases such as "tidak puas" are included so negation is captured. This is equivalent to the tokenising stage of text-processing operators.
Select Attributes with the regular expression w_.* keeps only the word columns and the label; the original text is removed.
Generate TFIDF gives higher weight to words that appear in fewer reviews. Replace Missing Values fills 0 for reviews with no keyword.
Split Data (stratified, 70:30) then Multiply feeds the training data to both models and the word weighting, and the test data to both evaluations.
Naive Bayes uses Laplace correction; Support Vector Machine (LibSVM) uses a linear kernel with C = 10.
Apply Model and Performance (Binominal Classification) compute accuracy, precision, recall, F-measure and AUC for each model on the test data.
Weight by Information Gain scores every word. Weights to Data and Write CSV save it so it can be sorted and read in Excel.
ulasan_produk.csv
2,400 rows · 3 columns
2,400 fictional product reviews in Indonesian (short sentences, some mixed and some with negation) with a positive or negative sentiment label.
Ready-to-run RapidMiner process (.rmp); CSV dataset with data dictionary; Practice guide and exercises (PDF/MD); Step-by-step video tutorial.
Not necessarily. The project is rated "advanced", and every step of the process is explained so you can follow how the data becomes a result.
Yes. Swap the dataset in the first step and adjust the attribute roles; the rest of the process stays the same. If you would like us to adapt it for you, message us on WhatsApp.
The project costs US$31. Use the "Buy via WhatsApp" button on this page; the project files are delivered after payment is confirmed.
Found a problem or have a suggestion for this project? Tell us on WhatsApp
Back to RapidMiner portfolio