Please enable JavaScript.

Coggle requires JavaScript to display documents.

QUESTION - Coggle Diagram

- - - - Have Order
        
        Examples:
        Low /Medium/High
        Education classes (1-10)
        Good /Bad/Ok
        University course
      - Encoding
        
        Label Encoding
        Target guided ordinal encoding
    - - No Order
        
        Examples:
        Pen/Pencil/Eraser
        Male/Female
        Cat/Dog/Deer
        Political Parties
        Eye Colors
      - Encoding
        
        One Hot Encoding
        One Hot Encoding with many
        Mean Encoding
  - - - Examples:
        Foot Size
        Height
        Temperature
      - 1,1.1,1.2 etc.
      - no.of variable
        
        Single Variable
        
        Histogram
        
        Density Plot
        
        Box-Whisker
        
        Two Variables
        
        Scatter Plot
      - fit in the interval
    - - Examples:
        Coin Toss
        Dice Throw
        Shoe Size
        Doctor Visits
        No. of cousin
      - 1,2,3 etc
      - No of Variables
        
        Single Variable
        
        Bar Chart
        
        Two Variables
        
        Mosaic Plot
        
        Stacked bar Chart
    - - Box Whisker
      - Pie Chart
      - Dot Chart
      - Strip Chart
- - - - Type of Biases - Notes
        
        reponse bias ----Broken (theramoter broken), embrassing(question asked), baised wording
        
        Undercoverage Bias--->Selection guys, Voluntary response sampling, conveniece sampling
        
        Non responsive bias---> lot of guys did not repond the survey.
        
        Find direction of Bias where we are getting skewed.
        Overestimating, underestimate etc.
        
        How to avoid Bias:
        Use sample properly as mentioned here
      - Low Bias
        
        Close to Truth
        we want low bias and low variance in model
      - High Bias
        
        Away from truth
      - Bias - average distance between predictions and the truth.
        Variance - Difference between predictions from different models.
        optimization of one will deoptimization other one.