Hands-On Industrial Internet of Things

Steps 0 and 1 – problem statement and the dataset

The dataset can be downloaded from https://ti.arc.nasa.gov/opensource/projects/. Click on Multiple Kernel Anomaly Detection (MKAD) Algorithm. For this simplified exercise, we will use only the first file. The challenge proposed by NASA is to detect anomalies .

The data is provided in a compressed CSV format where the 16 columns are Time, Altitude, AirSpeed, Landing_Gear, Thrust_Rev, Flaps, and a further 10 generic parameters (Param1_1, Param1_2, ..., Param2, Param3_1, ..., Param4).