Python Forum

Full Version: Recommendations for ML libraries for time-series forecast
You're currently viewing a stripped down version of our content. View the full version with proper formatting.
Hi fellow data scientists,

I am new to ML and python and after some research I am looking for some hints or guidance:

I have 3 kinds of problems(classification, clustering and time-series forecast). For classification and clustering I have decided to give scikit-learn over tl/keras a chance. The reason for that was it seems a bit easier for beginner and my problems don't require extraordinary tuning. As scikit-learn does not seem to be adequate for time-series forecast, I came across two models (ARIMA and LTSM, I am sure, there are more) and statsmodels (https://www.statsmodels.org/stable/gettingstarted.html) and sktime (https://www.sktime.org/en/latest/index.html).

Are there any recommendations on both, models and libraries, to beginn with as a beginner?

My time-series problems are the following:
1) Based on previous data, forecast on a charging point at connecting, when the charging will be finished/80%
2) Forcast on provided energy of multiple multiple charging stations

Thanks in advance, I'd appreciate a hint. If there's any info missing, please tell me.