Regarding deep learning: In an environment enabled with sophisticated ''Internet of Things'' solutions (e.g. a sensor-equipped automated warehouse) where a vector of hundreds of measurements is collected periodically (e.g. once per hour), how can an unsupervised deep learning solution be designed for anomaly detection based on e.g. years of monitoring? Note: *anomaly detection means here detecting that something in the controlled environment is not going right, is not as usual, is not as it should be. Note: **for the sake of simplicity, skip time dependency between measurements (which could be plausible instead) and only require the definition and discussion of an algorithm (i.e. conceptual steps, choices, motivation) implementing an analytics pipeline to pursue the given goal.
Regarding deep learning:
In an environment enabled with sophisticated ''Internet of Things'' solutions (e.g. a sensor-equipped automated warehouse) where a
Note: *anomaly detection means here detecting that something in the controlled environment is not going right, is not as usual, is not as it should be.
Note: **for the sake of simplicity, skip time dependency between measurements (which could be plausible instead) and only require the definition and discussion of an
Trending now
This is a popular solution!
Step by step
Solved in 2 steps
