THE RESULT
What the data shows
The largest computed metric is 0.7205 for 154.0; the smallest is 0.3702 for 147.0. Metric: review_score (Relative anomaly score).
There are no anomaly labels. The review fraction is a workflow assumption, not a measured error rate.
| row id | fresh | milk | grocery | frozen | detergents paper | delicassen | review score |
|---|---|---|---|---|---|---|---|
| 154 | 622 | 55 | 137 | 75 | 7 | 8 | 0.7205 |
| 338 | 3 | 333 | 7,021 | 15,601 | 15 | 550 | 0.6651 |
| 65 | 85 | 20,959 | 45,828 | 36 | 24,231 | 1,423 | 0.6293 |
| 75 | 20,398 | 1,137 | 3 | 4,407 | 3 | 975 | 0.6276 |
| 183 | 36,847 | 43,950 | 20,170 | 36,534 | 239 | 47,943 | 0.6166 |
| 161 | 12,434 | 540 | 283 | 1,092 | 3 | 2,233 | 0.613 |
| 142 | 37,036 | 7,152 | 8,253 | 2,995 | 20 | 3 | 0.6106 |
| 47 | 44,466 | 54,259 | 55,571 | 7,782 | 24,171 | 6,465 | 0.6087 |
| 356 | 22,686 | 134 | 218 | 3,157 | 9 | 548 | 0.6051 |
| 128 | 140 | 8,847 | 3,823 | 142 | 1,062 | 3 | 0.599 |
| 85 | 16,117 | 46,197 | 92,780 | 1,026 | 40,827 | 2,944 | 0.5988 |
| 289 | 42,786 | 286 | 471 | 1,388 | 32 | 22 | 0.5986 |
THE METHOD
From source to answer
Isolation Forest on log-transformed standardized category spending, with a declared 5% review fraction.
The Python source exposes this study’s transformations. The complete project download includes shared preparation and evaluation routines.
THE NEXT DECISION
What follows from the finding
Review atypical profiles for data quality or legitimate specialized purchasing; do not label them fraudulent.
Where the conclusion stops
Annual customer spending uses source monetary units, not an assumed currency. There are no margins, transactions or dates. Customer segments are descriptive and do not establish promotion response.
Related studies may reuse observations or holdouts. These are historical analyses; associations and backtests do not demonstrate commercial impact. Further model tuning needs new, untouched evaluation data.