Icon

A) Case - Fraud detection

Model Training and Evaluation
Exploration and Preparation
Load data
Deployment
CONTEXT: We are considering a large banking company operating in Europe that is facing issues related to fraud against its customers. DATASET: Consider the datasets 'fraud_dev.csv' and 'fraud_scoring.csv', which include behavioural data of customers of the bank. Some customers included in the dataset are related to fraudulent activity on their bank account (consider the variable FLG_FRAUD, where 1="fraud" and 0="otherwise"). GOAL: The bank's objective is to predict possible frauds before they occur and to take appropriate countermeasures to stop the most likely frauds. Can a ML model effectively predict potential frauds? MISCLASSIFICATION COSTS: The average cost of a fraud incident is $50,000, while the estimated investigation cost of identifying a potential fraud (that later turns out to be a false positive) is $400.
Column Appender
Import scoringdataset
CSV Reader
Data Cleaning and Feature Engineering
Import dev dataset
CSV Reader
Risk Classes
Pie Chart
Logistic Regression
Model 1
Train 70% Test 30%
Table Partitioner
Random Forest
Model 2
Color Manager
Balancing
Model evaluation
Clients list export
Excel Writer
Data prep
Random Forest Predictor
Exploratory Data Analysis
Risk classes
Numeric Binner

Nodes

Extensions

Links

Download

To use this workflow in KNIME, download it from the below URL and open it in KNIME:

Download Workflow
Created by: Vale
Created at: 2017-04-25
On NodePit since: 2026-03-04
Last update: 2026-03-04
Created with KNIME version: v5.8.2

Deploy, schedule, execute, and monitor your KNIME workflows locally, in the cloud or on-premises – with our brand new NodePit Runner.

Try NodePit Runner!