{

"SUBCATEGORY_ID": "01",

"SUBCATEGORYCONFIDENCE": 90,

"emoji": "📺",

The human is explaining a problem to the AI. The problem is an anomaly detection machine learning 🎉 problem. The dataset consists of key performance indicators of various aspects of different companies, and the goal is to detect 🎉 anomalous values in these KPIs. The judgement of anomalous values should be made by comparing similar companies, with similarity defined 🎉 by company size, revenue, geographic region, and industry. Anomalies should be looked for both within a single time period and 🎉 over time. The model's predictions should be explainable.