Upload dataset
CSV, TSV, TXT, DAT, XLS, XLSX, or XLSM. The first row must contain column names.
STATISTICAL & MACHINE LEARNING
Train and evaluate models directly from your dataset.
Load a tabular dataset, configure preprocessing and validation, train models, evaluate results, and export reproducible reports.
Ready
Machine-learning task
Choose how the dataset will be modeled.
Feature columns
Numeric columns are selected automatically. You can exclude columns before training.
This version trains on numeric predictors. Categorical target columns are supported for classification; non-numeric predictor columns are excluded to avoid hidden encoding and data leakage.
Missing values
Choose the rule used inside every training fold.
Transform & scaler
Transformation is applied first, then scaling is fitted using training data only.
Dataset
→Mean imputation
→No transform
→StandardScaler
→ML
Principal Component Analysis (PCA)
Unsupervised projection of the selected feature matrix before machine learning.
PCA has not been calculated.
Linear Discriminant Analysis (LDA)
Supervised projection for classification. Class labels are used only in this LDA analysis.
LDA has not been calculated.
PCA and LDA are diagnostic visualizations. Machine-learning training continues to use the selected original features unless you explicitly change the feature selection.
Model
Select an algorithm and its main parameters.
Validation
The test set stays independent from cross-validation.
Preprocessing parameters are fitted separately inside each training fold to reduce leakage.
Hyperparameter Tuning
Optional model selection using cross-validation on the training partition only.
Search spaceComma-separated candidate values. Grid Search evaluates every combination; Random Search samples combinations.
0 combinationsTuning runs only inside the training partition. The independent test set remains untouched until the final model evaluation.
Ready to train
Review the pipeline then train the model. Hyperparameter tuning is optional.
No model trained yet.Complete steps 1–4 and train a model.
MODEL EVALUATION
Training result
Evaluation metrics
Complete aggregate metrics for the independent test set.
Evaluation plot
Diagnostic plot
Detailed evaluation
Cross-validation
MODEL RUNTIME
Load Model & Prediksi
Muat model STASTUC yang telah disimpan, masukkan data baru, lalu jalankan prediksi tanpa melakukan training ulang.
1. Load Model
Gunakan file model JSON yang diekspor dari STASTUC.
Pilih model STASTUC
JSON model • drag & drop atau pilih file
Belum ada model yang dimuat.
2. Input Data
Masukkan satu sampel secara manual atau unggah dataset untuk prediksi batch.
Load model terlebih dahulu. Kolom input akan dibuat otomatis sesuai fitur model.
Dataset prediksiCSV, TSV, TXT, DAT, XLS, XLSX, atau XLSM. Nama kolom fitur harus sesuai model.
3. Hasil Prediksi
Hasil menggunakan preprocessing dan urutan fitur yang tersimpan pada model.
Hasil prediksi akan muncul di sini.