Statistical estimation of the prediction accuracy of breast cancer malignancy diagnosis in order to prevent the disease

Breast cancer represents a formidable peril to the female populace on a worldwide level and most common causes of death among women. With early detection of breast cancer and timely treatment, the chances of survival increase. Early detection of cancer, which usually results in reducing the extent of damage, less extensive treatment and better outcomes. The present study aims to statistically estimate the accuracy of predicting the malignancy of breast cancer in women and to prevent disease. The present study is a descriptive-analytical study with a sample size of 569 women with benign and malignant breast cancer with 32 features on the breast cancer data set of the UCI database. In the work, the implementation was made in Python, using librarie (Keras). After data normalization, a neural network model based on perceptron structure and Keras library was used to estimate the accuracy of breast tumor malignancy. The results of the present study show that after pre-processing the disease dataset, the accuracy of the proposed model for the training and test data was 0.97 and 0.96, respectively, which is considered high accuracy for this dataset. A neural network was able to discriminate with high accuracy the two separable sets discriminating the benign or malignant tumor patients. The findings of this study will help in the early detection of malignant tumors in patients with breast cancer and effective decision-making for its treatment.

Statistical estimation of the prediction accuracy of breast cancer malignancy diagnosis in order to prevent the disease Keywords:

Accuracy , Breast cancer , Statistcs , Keras.