Issue
I am trying to do a university project, and my professor asked me to do a Neural Network to predict servers' behavior in this CSV data. https://drive.google.com/file/d/1Su8CnliKcclYxmSOlEg7zu_73X2WO9J4/view?usp=sharing
nodeid = id of a server
node_cpu_usage = server cpu usage
node_memory_usage = server memory usage
And the task is to train a model on this server's CPU and memory usage to predict how they will behave in the future. What is your suggestion?
Solution
This falls under time anomaly detection using 2 input variables: (CPU, & memory). In NN world, the time series recurrence is trained using LSTM. If you Google with search terms time anomaly detection using LSTM
or multivariate anomaly detection
, you will get lots of academic work about this type of implementations, and some sample reference implementations. Not all these implementations use neural network; but should give you an idea about exploring and pre-processing your data.
Some references for help:
- https://becominghuman.ai/time-series-and-how-to-detect-anomalies-in-them-part-ii-bde9e69d0aaf
- https://datascience.stackexchange.com/a/77872/38183
Answered By - Azhar Khan Answer Checked By - Marilyn (PHPFixing Volunteer)
0 Comments:
Post a Comment
Note: Only a member of this blog may post a comment.