A Comparative Study of Gene Expression Data-Based Intelligent Methods for Cancer Subtype Detection

Cancer is one of the major causes of human death. Hence, diagnosis and treatment of cancer are considered as essential tasks. The targeted treatment of cancer disease varies according to the subtype of cancer. Due to this reason, researchers are involved in the development of new methods to classify the cancer subtypes accurately. The genetic level study plays a vital role in the treatment as well as classification of cancer types. In recent years, a large volume of gene expression data is accessible to predict the cancer subtypes. Numerous intelligent systems were developed to classify the cancer types using gene expression data. In this paper, a comparative study of gene data analysis-based intelligent systems used for cancer detection is presented. The important issues in gene data analysis-based cancer detection were detailed before introducing artificial intelligence (AI)-based classification systems. This study is intensive to the cancer detection approaches based on machine learning (ML) and deep learning (DL) techniques. The benchmark gene expression datasets used in cancer subtypes classification were initially detailed, and the performance of different intelligent cancer subtype classification methods was compared to judge the classification efficiency. Furthermore, a deep structured reinforcement learning (RL)-based cancer detection method was also proposed in this research article to progress the classification performance, whereas the RL techniques are used for cancer detection.KeywordsCancerGene expression dataClassificationArtificial intelligenceMachine learningDeep learningReinforcement learning