Option challenge : Options Explorer
Imagine that you started working in a company that collects and analyzes high-frequency grinding machine sensor data to detect machine anomalies. You are tasked to find an optimal model for your dataset under certain limitations and preferences.
The intent can be described as anomaly detection with the general method of analysis multiclass classification. Preferably, your model should fulfill the following soft constraints: (1) accuracy of at least 0.9, (2) precision of at least 0.9, (3) recall of at least 0.8, (4) F1 score of at least 0.85, and (5) the processing unit (PU) to be GPU.
Background
the dataset UC5_expanded.csv from OptionsExplorer materials
the Options Explorer GUI http://194.249.3.27:3000/
Select the dataset UC5_expanded.csv and upload it.
In Required Parameters, choose the domain of knowledge as the manufacturing industry. In Optional parameters, specify the intent as anomaly detection and the general method of analysis as multiclass classification. They will serve as hard constraints.
Choose the variables from the drop-down menu and specify thresholds or categories for each of them.
- (1) accuracy ≥ 0.9,
- (2) precision ≥ 0.9,
- (3) recall ≥ 0.8,
- (4) f1_score ≥ 0.85, and
- (5) pu = GPU.