Microarray Data Classification to Detect Cancer Cells by Using Discrete Wavelet Transform and Combining Classifiers Methods

Cancer has become a disease with the highest worldwide mortality rate, reaching 9.6 million occurrences in the year 2018. Researchers are using microarray data to observe the level of cancer expression gene. However, microarray data have huge data attribute and it causes curse of dimensionality. Thus, data processing takes a longer time. Dimensional reduction technique by using Discrete Wavelet Transform is being used in this research to solve these problems. The dimensional reduction process is utilizing the family daubechies4. Then, a method between K-nearest Neighbor and Support Vector Machines is chosen based on the neighbor similarity during the data classification process. Therefore, the created system could produce 95% classification accuracy for colon cancer's data, 88.88% for breast cancer's data, 87.16% for lung cancer's data, and 100% for ovarian cancer's data.