fastlane is winner.

No

Detecting a behavior anomaly The dataset below contains a location timeserie of a person living alone in their appartment. The data indicates in which location/room he was at which point in time. This data has been collected by sensors installed in the home. This person had a health incident in the night from 2019-09-18 to 2019-09-19, probably at 3:00 or 10:00 which shows in a drastic change in location behavior and resulted in the person going to hospital. By the sensor setup in the home, the location entrance and livingroom are really one single bigger room. Merge them. Explore the data and give an overview of key metrics (graphically and quantitatively) Can you say something about the living routines of the person? Propose one or more methods to detect the incident in "real time" by analyzing the location data. Real-time means, that while time passes more and more of the data gets "known" to your detection method. It can trigger as soon as the incident is detected, an action can be triggered. We are interested in understanding how you proceed in analyzing this case. Show your thought process What methods did you try and why What are their strength and weaknesses of the approaches. Are they robust and generalizable to other users? How do you test your code for correctness? We are looking forward to your propositions! PS: You are free to use other Python libraries as desired. Please return your Notebook as an answer.

https://githubhelp/deepsparkchaker/anomaly-behavior-detection

10650

4

0

1

2020-04-22

1

1

0

deepsparkchaker