Join the conversation
![](https://codanics.com/wp-content/uploads/2024/04/LinkedIn-profile-picture.jpg)
I learned about Data Transformation.
--> These are 3 Scalars for Linear Data
# Standardization or mean Removal:
1:Standard Scalar [-3,3] (used for models which deal with -ve values as well)
# Scaling to a known range:
2: Min-Max Scaling [1,0] (used for models which deal with +ve values only)
3: Max-Absolute Scaling [-1,1] (used for models which deal with -ve values as well)
--> If the data has non-parametric distribution i.e. the data is not normally distributed, then we use Quantile Transformer.We also use Box-Cox, Yeo-Johnson, Quantile transformers.
Reply
![](https://codanics.com/wp-content/uploads/2024/04/IMG_16954467347071342.jpg)
Done
Reply
![](https://codanics.com/wp-content/uploads/2024/04/IMG_20240410_175137-1-scaled.jpg)
Done
Reply
![](https://codanics.com/wp-content/uploads/2024/05/My_profile_pic.jpg)
I learned in this lecture 1: Standard Scalar ( -3 —------- +3 ) ( handle negative value ) 2: Min-Max Scalar ( 0 —--------1) 3: Max-Absolute Scalar ( -1 —-------- +1 ).
Reply
![](https://codanics.com/wp-content/uploads/2023/10/x8h6vl_1.jpg)
jazakumulah kharn
![](https://codanics.com/wp-content/uploads/2023/10/e24e3cc0-f9cd-49cc-9e5e-e7fea619bd42.jpg)
I learned 1: Standard Scalar ( -3 —------- +3 ) ( handle negative value ) 2: Min-Max Scalar ( 0 —--------1) 3: Max-Absolute Scalar
( -1 —-------- +1 ).
Reply
![](https://codanics.com/wp-content/uploads/2023/10/9dd24f5a-b137-440a-baba-1855925152a0.jpg)
AOA, I learned in this lecture about transformation and its types.
TYPES OF TRANSFORMATION1: Standard Scalar ( -3 —------- +3 ) ( handle negative value )
2: Min-Max Scalar ( 0 —--------1)
3: Max-Absolute Scalar ( -1 —-------- +1 ) ( handle negative value )I also learned about non-parametric distribution, which is a QUANTILE transformation (converting the data in uniform), and another one is MAPPING (converting the data to a Gaussian distribution). ALLAH PAK aap ko sahat o aafiat wali lambi umar ata kray aor ap ko dono jahan ki bhalian naseeb farmaey Ameen.
Reply
![](https://codanics.com/wp-content/uploads/2024/02/2.jpg)
I learned standardization steps (i.e. standard scaling, min-max scaling, and max-abs scaling). If algorithms deal with negative values then we can use standard and max-abs scaling otherwise min-max scaling
Reply