Great to hear!

Thank you! 😃

Glad you enjoyed it

You use Convolutions when you have images to extract features, since you already have all the features extracted, and the data has no images you don't need convolutional layers

Same it will be great if you add some tutorial on breast cancer detection on mammography images

There are NA columuns after reading the csv file. I fixed the problem by removing these columns using "df.drop(df.filter(regex="Unname"),axis=1, inplace=True)" line of code after the reading command. I hope this would help you.

@@mistirmulla thanks a lot!

for all who will have the same probleme , You replace the code lines with this : df.rename(columns={'diagnosis': 'Label'}, inplace=True, errors='raise') print(df.dtypes)

Structured data is part of experimental results, even for image processing. You never make a decision based on one image or one experiment. You collect enough data to understand the trends. Therefore, it is very important for researchers to learn these concepts.

Sorry, I did not understand the question. In this example the data has been provided in a structured way where the features are pre-engineered (provided as attributes). If you can gain insights directly from the attributes then you do not need deep learning. The mapping from the attributes to output need to be learned via training. For structured data, especially if you only use dense networks you only train the fully connected neurons. For images, you can supply pre-engineered features that you generate by applying various filters. But that approach is not ideal as you do not know which filter works well for your specific case. Therefore, you use trainable convolutional filters to extract features and constantly learn as you train the model. This feature information is then fed into dense network for classification. In summary, features can be pre-engineered or learnt from training. In both cases you can use traditional ML (e.g., Random Forest) or Deep learning to classify.

I did a quick google search for the data set and I found it on kaggle. I recommend doing the same for any data set, you will usually find it online. Here is the direct link from Kaggle. www.kaggle.com/uciml/breast-cancer-wisconsin-data

Because I want to create a project with own dataset.🙏

For images, you can apply many digital filters to extract features. For structured data you can combine various attributes and this field is called feature engineering. It requires domain knowledge as you cannot just randomly start combining various attributes. Usually, deep learning works fine with structured data using the attributes provided as it fits non-linear models.

@@ZEISS_arivis sir you have a breast cancer image dataset link not histopathology image? If you have so please share with me. 🙏🙏🙏🙏

There are NA columuns after reading the csv file.So we need to remove these what called Unnamed columns. I fixed the problem by removing these columns using "df.drop(df.filter(regex="Unname"),axis=1, inplace=True)" line of code after the reading command. I hope this would help you.

@@mistirmulla hall ettim Mustafa bey tesekkur ederim.

@@mistirmulla thanks a lot!