There is irregularity in time interval of each sensor of quantitative Properties. Can you help with that?
Hi Sourabh. That is part of the problem you have to solve. The data is irregular and not all sensors are active. You could consider different statistical techniques (linear interpolations, kalman filter, etc) to fill in the gaps. You will be scored for novelty.
yes that’s right missing value imputation is fine but timestamp values are not fix like suppose for example values are coming for random time intervals ain’t it should be hourly or half hourly or for every 15 mins?
The period of sampling is once every 15 minutes, but due to various issues like sensor not giving data, server not responding, etc, there might be some randomness in the interval. As such, most of the _MAX and _MIN values are 24 hour aggregates, which can be seen from the description of a quantitativeAttribute.
Let me know if this answers your query.
Yes thank you very much. I thought their will be NaN values for for such issues. So first we need to find which timestamps are missing and then add NaN values their and then impute and then do all stuff. Cool thanks