AI-based detection of Breast Cancer tumor cell using Machine Learning Models
It is not mandatory for you to be a Doctor or Biology student anyone can provide a solution under Health Care. It doesn’t mean that you should have any biological background to consider any problem statement under Health care. The important requirement is Requirement Gathering and impact analysis by having this two as a key tool we can provide solution on any field.
So Let’s get started.
INTRODUCTION
The diagnosis of breast cancer involves the classification of its tumor cells into a malignant and benign group into the subject of much research. The future prediction of benign or malignant of a patient is more complex. Further classification of benign tumors can prevent patients from undergoing unnecessary treatments.
Because of its unique advantages in critical features detection from complex BC datasets, machine learning (ML) is widely recognized as a choice in Breast Cancer tumor classification.
UNDERSTANDING THE DATA SET:
Breast cancer occurs in various factor depending upon the patient(host). However, most case of breast cancer cannot be linked to a specific cause.
FACTORS:
1)Age.
2)Genetic factors.
3)Family history of breast cancer.
4)Childbearing & Menstrual history.
PREPROCESSING THE DATA SET:
So from the list of features given below for a numerical data set with proper feature engineering and preprocessing of data we can select appropriate algorithm to classify the target variable (Benign or Malignant). For example, we can use classification algorithms to classify a mail is a SPAM or NOT SPAM.
So Let’s head to our topic.
The following is a sample list of features we could consider when building a classification model.
1)Diagnosis (M = malignant, b-benign)
2)Radius(mean distance from center to points on the perimeter)
3)Texture(standard deviation of grayscale values)
4)Perimeter
5)Area
6)Smoothness(variance in radius length)
7)The compactness of the cell
8)Concavity (rate of change of the cell from one to another)
9)The symmetry of the cell
10)Fractal dimension (how complicated the cell grows)
SPLITTING THE DATA SET:
Once we have done with preprocessing of data we need to split the data into training and test data.
Once we have done with the training and testing of data using confusion matrix we need to validate the model.
CHOOSING THE MODEL:
To achieve this it is required to use various machine learning classification algorithm to fit the model that uses testing and training of data set for a better score of accuracy in the outcome.
PUBLISHING THE MODEL IN WEB SERVICE:
After developing the model we are in the final stage, that is to deploy the model into a web service like AWS and to make it easier for people to do their prediction with the classification model.
CONCLUSION:
So using this information, Doctors can start treatment for the patients with benign tumor to prevent the cell growth to Malignant in future.
