> symptons

I have chest pain

> age?

I'm fifty yesrs old

> gender?

female(LGBTQ+)

> DO u have a ct scan

upload scan.jpg

Keyword extraction demo:https://www.textrazor.com/demo

Keyword blog:https://monkeylearn.com/keyword-extraction/

The dataset is downloaded from lung cancer dataset. This dataset is preprocessed and it is in a very good shape in terms of tidiness. Loaded from the csv file, the first ten rows of the dataset is shown below.

After removing the patient_id column, which is irrelevant to our training, the dataset is 1000 samples by 23 features. The model takes in the first 23 columns as X, and the last column, level of severity for its training and testing.

mask = np.random.rand(n_sample) < 0.8
training_set = cleaned_df[mask]
testing_set = cleaned_df[~mask]

The data is loaded as a pandas DataFrame object, where we can easily split the dataset by 80% of training data and 20% of testing data.

testing_X = testing_set.loc[:, testing_set.columns != "level"]
testing_y = testing_set[["level"]]
nn.score(testing_X, testing_y)

The model is trained on a training set of 797 samples and tested on a testing set of 203 samples, the total score is 1.0, which means it predicted all of the samples in the testing set correctly.

nn.predict(scan.jpg)

return result to chatbot

Models for breast cancer detection:

code:https://github.com/lishen/end2end-all-conv paper:https://paperswithcode.com/paper/deep-learning-to-improve-breast-cancer-early

• inform results
• give expectation life time, probability
• provide advice
• provide medication advice
• give prescription

• recognition
• CNN classification
• regression

• Flask API as backend
• Frontend should be a WebApp