Analysis and Prediction of Cancer using Genome by Applying Data Mining Algorithms

Abstract

The complete set of DNA is represented by a sequence called genome which has all genes. To build and maintain any organism, every genome contains the entirety of the data required. In human body, more than 300 crores DNA base pairs are maintained with all cells and their nucleus. The complete study of genome is called genomics. newlineInformation mining is the course toward finding structures in huge educational collections including techniques at the convergence motivation behind bits of information and database frameworks. It is an interdisciplinary sub field of programming building and estimations with a general goal to discard information (with sharp methods) from an educational assortment and change the information into a coherent organization for extra use. There are two functionalities-arrangement and grouping which can be applied on information to digest information from the huge dataset. newlineIn this research work, genome study is done and on that study we have applied few data mining techniques like supervised and unsupervised learning on cancer data sets. The data set is available on the website of Bioconductor and R packages are used for further analysis. The implementation detail is divided into four different parts. First part of research is based on classification where we have identified Leukaemia types by applying classification algorithms. The second part is to identify the subtypes of cancer using non supervised learning clustering. The next parts are focused on different types of clustering methods which can be applied on genome study and also perform some fusion of clusters. newlineAs a part of pre-processing techniques, outliers needs to be removed so cleaning of data, transformation and reduction of data is applied. In the existing algorithms of data mining, there are many short comings. In this, an attempt is to make to overcome the disadvantages offered by existing algorithms by applying some mix and modified approaches and try to improve the prediction of the cancer disease from the genomics. newline