Can we input confusion matrix data to sklearn just for plotting confusion matrix?

1

I have a model in matlab , which I am not able to regenerate in Python. My question is can I just plot the confusion matrix in python, using any lib like sklearn or so by just inputting the confusion matrix.
Plz help

2

You just need predicted values and expected values to have your confusion matrix, with sklearn.metrics.confusion_matrix for example.

from sklearn.metrics import confusion_matrix
import seaborn as sns
import matplotlib.pyplot as plt

def plot_conf(y_true, y_pred, label : str = "", figsize=(7,4)) :
    cm = confusion_matrix(y_true, y_pred)
    fig, ax = plt.subplots(figsize=figsize)         # Sample figsize in inches
    fig.suptitle("confusion matrix %s"%label)
    f = sns.heatmap(cm, annot=True, linewidths=.5, ax=ax)

So you can for example find a way to store, from matlab, the latter in a file that you will read afterwards in python for your computations : for example for a vector y, a text file containing on each line an element of y, so that the reconstruction in python will be easy.

Sorry, I don’t master matlab, this is just the idea that comes directly to my mind.

1 Like
3

Sir I am not able to reproduce the predicted and true values here in python. Let me see, if i can do anything for that. But thanks for reply in my other code here in python I got this
image
Sir can I normalize it to find accuracy in percentage, like not in number.

Regards

5

You can just follow this link. You need to add this line

cmn = cm.astype('float') / cm.sum(axis=1)[:, np.newaxis]
1 Like
6

Thanks mam @krishna511

1 Like