sktime - A Unified Toolbox for ML with Time Series - Markus Löning | PyData Global 2021

Поділитися
Вставка
  • Опубліковано 13 гру 2024

КОМЕНТАРІ • 10

  • @utkar1
    @utkar1 2 роки тому +13

    Time stamps (WIP)
    @00:00 - Introduction
    @02:21 - Agenda
    @04:51 - Github Repository
    @05:58 - Intro to sklearn
    @11:42 - Intro to ML with Time Series
    @18:25 - Creating a unified framework
    @21:15 - Forecasting

    • @MayuriK_it
      @MayuriK_it 2 роки тому +1

      52:38 - Univariate forecasting with exogenous variables
      56:20 - Multivariate forecasting
      1:04:20 - Time series classification
      01:14:20 - Time series regression
      01:16:40 - Multivariate time series classification

  • @cHrIs16249
    @cHrIs16249 2 роки тому +1

    01:14:20 Time series regression

  • @davidk.9560
    @davidk.9560 2 роки тому +3

    @1:04:20 - Classification

  • @RAHUDAS
    @RAHUDAS Рік тому +2

    "Thank you for the detailed explanation. However, I would like to assess the performance of my model on the training data. Specifically, I want to plot the model's response against the actual training data. Could you please provide guidance on how to accomplish this using the sktime library?"

  • @luiscunha844
    @luiscunha844 2 роки тому +1

    Is there some simple way for me to turn a normal csv file (ex Date Open High Low Close) to the required format for multivariate timeseries?

  • @zhaohaiding9220
    @zhaohaiding9220 Рік тому +1

    讲得很好!

  • @giorgiotelarico7283
    @giorgiotelarico7283 Рік тому

    yes without shifting the target...

  • @viniciusdesouzamenezes6594
    @viniciusdesouzamenezes6594 Рік тому

    hi, i`m Vinicius, I would apreciate a personal suport to sktime aplication, is there someone

  • @AIandVisuals
    @AIandVisuals Рік тому +1

    You run a mean squared error on a classifier´s predictions??? This makes no sense. You should have used accuracy or a confusion table instead. This is basic.