New Assessment

Configure your model pipeline and upload a transaction dataset to begin detection.

Assessment Configuration

Select a model and decision parameters, then upload your CSV

1 Model Pipeline

Apex 1.0

Ensemble oracle — combines RF, XGBoost and LightGBM with confidence weighting

RECOMMENDED

Core XGBoost

Gradient boosted trees — high precision, full PR-curve threshold control

PR CONTROL

Core LightGBM

Leaf-wise growth — fast inference, full PR-curve threshold control

PR CONTROL

Core Random Forest

Decision tree ensemble — stable baseline, full PR-curve threshold control

PR CONTROL

Confidence Threshold

50%

Transactions scoring ≥ 50% will be flagged as fraudulent

Precision–Recall Constraint

50%

Threshold0.5000

Est. Precision—

Est. Recall—

2 Upload Dataset

📄

Drop your .csv file here, or click to browse

Processing pipeline

Receiving and validating dataset

Normalizing features and engineering signals

Running inference pipeline

Serializing output to secure session

Assessment Results

Fraud predictions are ready. Preview below or open the full data wrangler.

Data Preview — first 10 rows