ray941216 / data-engineer-exam Goto Github PK

Hi @Ray941216 , I notice that train_model method uses the data in self.data.

If this model runs on the service, the time of the train_model method would get longer and longer?

Hi @Ray941216 , your solution is gorgeous, nice work.

If our customers give feedback to us they are suffering from alert fatigue after applying your solution. How you improve your model or adjust the parameters to reduce the model alarm?

Hi @Ray941216 , great work for use .groupby method to aggregate the file.

However, it's very slow. If we transformer a huge size file, it will spend a long time.

Could you try to use other methods from pandas to improve efficiency?

Is it possible that I get the anomaly points more than normal points?
As the exam3 request, I aggregate the data from milliseconds to minutes data point.

We get 386 minutes in training data. After predicting, I got 245 points anomaly points which is more than normal points. Is there any correct mapping for training data?