wiqaaas/youtube

Contains Notebook for my youtube channel videos

youtube

Contains Notebooks for my Youtube channel Videos.

youtube's Issues

only why you calculating differently intercept for Ridge_Regression_using_Gradient_Descent

great code
https://github.com/wiqaaas/youtube/blob/master/Machine_Learning_from_Scratch/Ridge_Regression/Ridge_Regression_using_Gradient_Descent.ipynb
and video
https://www.youtube.com/watch?v=1j6DClKQjZY&ab_channel=wiqaaas

only why you calculating differently intercept

if feature_is_constant == True:
    derivative = 2 * np.dot(errors, feature)

since in
https://github.com/chasinginfinity/ml-from-scratch/blob/master/02%20Linear%20Regression%20using%20Gradient%20Descent/Linear%20Regression%20using%20Gradient%20Descent.ipynb
and
https://www.youtube.com/watch?v=4PHI11lX11I&t=536s&ab_channel=AdarshMenon

it calculated as
D_c = (-2/n) * sum(Y - Y_pred) # Derivative wrt c

Data Set

How can I get the dataset ?

I'm getting an error using example_ANFIS.py

484 RMSE is: 0.20023371637733992
485 RMSE is: 0.20023371637733992
486 RMSE is: 0.20023371637733992
487 RMSE is: 0.20023371637733992

...499 RMSE is: 0.20023371637733992

After running the file, I'm getting the following:

ValueError Traceback (most recent call last)
~\AppData\Local\Temp\ipykernel_9540\3618313043.py in ?()
8 bestParam, bestModel = E_Anfis.fit(X_train,y_train,X_test,y_test,optimize_test_data=True)
9
10 from scipy.stats import pearsonr
11 pred_train = E_Anfis.predict(X_train,bestParam,bestModel)
---> 12 pearsonr(y_train,pred_train.reshape(-1,1))
13
14 pred_test = E_Anfis.predict(X_test,bestParam,bestModel)
15 pearsonr(y_test,pred_test.reshape(-1,1))

c:\Users\dabis\AppData\Local\Programs\Python\Python312\Lib\site-packages\scipy\stats_stats_py.py in ?(x, y, alternative, method)
4804 msg = ("An input array is nearly constant; the computed "
4805 "correlation coefficient may be inaccurate.")
4806 warnings.warn(stats.NearConstantInputWarning(msg), stacklevel=2)
4807
-> 4808 r = np.dot(xm/normxm, ym/normym)
4809
4810 # Presumably, if abs(r) > 1, then it is only some small artifact of
4811 # floating point arithmetic.

ValueError: shapes (376,1) and (376,1) not aligned: 1 (dim 1) != 376 (dim 0)

Any ideas?

BraTs dataset config.json file

How to get config.json file??

why intercept is not calculated separately in Lasso_Regression_using_Coordinate_Descent.ipynb

may you clarify
why intercept is not calculated separately in
https://github.com/wiqaaas/youtube/blob/master/Machine_Learning_from_Scratch/Lasso_Regression/Lasso_Regression_using_Coordinate_Descent.ipynb

when it is calculated in
https://github.com/wiqaaas/youtube/blob/master/Machine_Learning_from_Scratch/Ridge_Regression/Ridge_Regression_using_Gradient_Descent.ipynb