STUDY / 01
TEMPLAR-Fraud lab
Independent implementation of a routed, verifier-grounded, calibrated and cost-sensitive fraud decision pipeline with synthetic demo data and a paper fidelity audit.
researchdemo readypublished · Discover Data 4:20
Business Analytics · AI · Research
Md Sultanul Arefin Sourav
Turning business data into decisions you can inspect.
I study business analytics at Trine University and explore how data science, machine learning, and automation can support better business decisions. Explore interactive systems for fraud review, operational planning, and evidence-based analytics.
- AFFILIATIONBusiness analytics research with an academic affiliation to Trine University.
- INTERESTSBusiness Analytics · Data Science · Data Analytics · Artificial Intelligence · Machine Learning
- MEMBERSHIPIIBA Member
Business Decision Network · interactive
Step a transaction from demand signal to reporting. Tap a node for its order, inventory, invoice and settlement state.
synthetic demo3D · WebGL
Nodes
Details
Nothing selected
Select a node in the scene, use the node buttons, or step a transaction. Escape clears the selection.
0 / 0 events
Research
Two studies, each with an evidence trail
STUDY / 02
Attention-QELM-GWO IIoT lab
Independent implementation of an attention encoder with a fixed-weight extreme learning machine head and grey wolf hyperparameter search for IIoT intrusion detection, with a fidelity audit.
researchdemo readyarticle in press (supplied version) · Discover Artificial Intelligence
Business demos
Systems you can step through
DEMO / 01
Business flow simulator
Deterministic discrete-event simulator of a fictional supplier-to-payment B2B workflow with idempotent events, review queues and policy comparison.
applieddemo ready
DEMO / 02
Cashflow invoice lab
Synthetic receivables lab: three-way matching, duplicate invoice checks, payment-delay baselines and a 13-week cash forecast with intervals.
applieddemo ready
DEMO / 03
Model monitoring lab
Drift, calibration and data-quality monitoring for prediction streams with a documented alert state machine and delayed-label handling.
applieddemo ready
How to read this site
Four evidence labels
paper-reported comes from a published table with its locator. reproduced would need a real-data run with a run id; nothing here carries it yet. demo is computed on an authored synthetic fixture. proposed has no value.
Fictional businesses
Every company, order, invoice and account in a demonstration is invented (Northwind Demo Supply and friends). No client, deployment or real dataset is implied.
Nothing moves on its own
A particle in the scene is an event from a replay file. Queue totals recompute from a demo trace. The paper's operating point is displayed, never edited.