ML Prediction of Breast Cancer Survival

Cancer is a process of abnormal cell growth which loses contact inhibition and cease proliferation and growth of normal cells when contacting each other. The old cells of the body grow out of control and do not die resulting in abnormal growth of cells forming a mass of tissue called Tumor while leukemia do not form tumors. Lung and prostate cancer is the deadliest cancer with the highest death rates. Pancreatic cancer is a silent killer.

Major Types of Cancer :

  • Leukemia
  • Sarcoma
  • Carcinoma
  • Melanoma
  • Lymphoma

Cancer are of various types depending upon the tissues and the organs they are infecting. Breast cancer is commonly seen in women and the symptoms are lumps, thickening and swelling of part of breast, redness or flaky skin, bloody discharge from nipple and irregular shape of breast. It can be treated by radiation, chemotherapy, hormone therapy and surgery based on the stage of cancer. 

Machine Learning Algorithm in Cancer Prediction
Machine Learning Algorithm in Cancer Prediction

Symptoms of cancer :

  • Hoarseness or Nagging cough
  • Change in Bowel 
  • Persistent lumps or swollen glands
  • Blood in Urine
  • Indigestion or difficulty in swallowing
  • Unexpected vaginal bleeding or discharge
  • Blood in stool 
  • Underweight loss, fever, night sweats
  • Continued itching in genital area

How does it occurs :

When normal breast cells begin to grow abnormally than healthy cells do, forming a lump or mass can easily spread to other parts of the body through lymph nodes. Through the ACS report the lumps can be easily seen. The division process may involve 28th to 30th cell division cycles to get detected. In malignant cancer the use of logistic regression machine learning technique has been implemented to detect the tumorous cells.

ML can be used to predict breast cancer in clinical practices with low to high accuracy. ML suggests an alternative prediction modelling approach to address the current situations and to improve the limitations. In order to predict the tumor cells deep learning techniques have been used with various activation functions like :

  • Tanh
  • Maxout
  • Exprectifier
  • Rectifier

SVM technique is suitable to identify the type of cancer if it is malignant or beginning based on tumor cell size. It is a branch of data science that uses an algorithm to sequence statistical data in a programmed manner.

ML Methods for automatic diagnosis of cancer

  • Novel Methods for image classification / segmentation
  • ML based Models for detection / prediction
  • Cancer image analysis
  • Data generation for cancer diagnosis
  • Mobile applications for detection / prediction

AI can be used in cancer predicting processes by applying various fields like ML, Neural networks, Evolutionary Computation, Vision, Speech Processing, Expert Systems, Natural Language Processing, Robotics and Planning. We can detect breast cancer through screening tests like mammograms, like an X-ray to detect cancer up to two years.

Cancer Survival Rate in Breast Cancer :

The average survival rate with non metastatic Invasive breast cancer is 90% (5-year Survival rate)
The average survival rate with non metastatic Invasive breast cancer is 84% (10-year Survival rate)
The average survival rate with non metastatic breast cancer is 99% (5-year Survival rate)