Recommendation System Attained by Analyzing and Predicting the Collaborative Data of Cancer

Abstract

Medical information volume and complexity have increased newlinesignificantly in the field of cancer care. But most of this data is found in newlineunstructured medical text, which makes typical analytical techniques newlinewhich are difficult to use. This work has focused on the analysis of clinical newlinetexts in the field related to cancer diagnosis. The research work focuses on newlinethree different objectives that support the identification and analysis of newlinebrain and kidney tumor from pathology report. newlineIn the first research objective Quantitative features from medical newlineimaging are extracted. Developing a unique histogram-based regionrelated newlinebrain tumor detection technique is the main objective. Utilizing newlineMRI images of kidney and brain, an automatic thresholding technique for newlinetumor detection and tumor segmentation is provided. The segmentation newlineresult processes for analyzing the image, that uses the intensity of the newlinepixels levels that are derived from every region of the histogram of an newlineimage. Based on the mean and the standard deviation of every area among newlinethe four sub-regions, the threshold is automatically chosen. The detailed newlinequantitative explanations, including statistical parameters, of the suggested newlinemethodology are provided which includes the outcomes produced by this newlinealgorithm are used for further investigation. The produced data will be newlineanalyzed and compared with a pathology report in order to diagnose and newlinepredict cancer. newlinev newlineIn the second research, objective is to provide a machine learningbased newlineclassification method that can classify benign and malignant newlinetumors at the patient level using clinical data and extracted quantitative newlinefeatures. Implementing a diagnostic model for benign and malignant newlinetumors for early prediction is done using machine learning algorithm. It newlineinvolves integrating the Cancer dataset of Brain and kidney with the newlinequantitative features for early detection. The first stage includes methods newlinelike data cleaning and tokenization, to clean and organise the newlineunstructured free-text data.