DBSCAN Clustering: Stop #4 on Your DIY Data Science Roadmap

Поделиться
HTML-код
  • Опубликовано: 28 дек 2024

Комментарии • 15

  • @DaveOnData
    @DaveOnData  2 месяца назад

    Save 20% off my machine learning online courses using code RUclips ⬇
    Cluster Analysis with Python online course:
    bit.ly/ClusterAnalysisWithPythonCourse
    My "Intro to Machine Learning" online course:
    bit.ly/TDWIIntroToML

  • @michaelt312
    @michaelt312 2 месяца назад +2

    Another happy Wednesday with another incredibly educational video from David.

    • @DaveOnData
      @DaveOnData  2 месяца назад +1

      As always, thank you so much for supporting the channel!

  • @davidghope1
    @davidghope1 2 месяца назад +1

    Another fantastic video, makes it so understandable and doable - just great😊

    • @DaveOnData
      @DaveOnData  2 месяца назад

      I'm very happy to hear that you are enjoying my crash courses!

  • @rauljg143
    @rauljg143 2 месяца назад +1

    Thank you for this. Your explanations and examples were really great. I'm doing this in R but it translated perfectly for me. 👏

    • @DaveOnData
      @DaveOnData  2 месяца назад

      Woohoo! Glad you found the crash course useful.

  • @VastCNC
    @VastCNC 2 месяца назад +1

    Thank you again.

    • @DaveOnData
      @DaveOnData  2 месяца назад

      My pleasure! I hope you find the video useful.

    • @VastCNC
      @VastCNC 2 месяца назад +2

      This actually peaks a particular interest. I’ve been needing a system to apply a k means centroid to a rectangular viewport (map), I’m only 10min in, but I’m thinking this may be a bit more useful to apply to rectangular map shapes

    • @DaveOnData
      @DaveOnData  2 месяца назад

      Yes, DBSCAN may prove more useful in this scenario.

  • @Bola382
    @Bola382 Месяц назад +1

    mixture models are cool too

    • @DaveOnData
      @DaveOnData  Месяц назад

      I haven't encountered mixture models as my business stakeholders/clients seek hard cluster assignments. Have you used them? I would be interested in any applications.

  • @KatiaBravo-d5b
    @KatiaBravo-d5b 25 дней назад +1

    Hi thanks for your explanation, Question? I was doing my hw, but my data did not form clusters. In the scatter plots it is showing separate each of one, quite disperse, also correlations are week, and when I try to apply DBSCAN or K-Means (curiosity) data is aligned vertically for each value in the axis X. ....What is wrong? What I can do? I normalize, standarized, scaled it ...thank you!

    • @DaveOnData
      @DaveOnData  25 дней назад

      That is strange. Have you optimized DBSCAN for your dataset? For example, optimizing eps?