A Novel Hybrid Dimensionality Reduction Technique for Cancer Classification from Microarray Data

Cancer classification using microarray data has become a critical area of research, given the complexity and high-dimensionality of genomic datasets. This paper proposes a novel hybrid dimensionality reduction technique for enhancing cancer classification accuracy by integrating two powerful methods: Principal Component Analysis (PCA) and Linear Discriminant Analysis (LDA). The proposed technique leverages PCA to capture the most significant global variance in the microarray data, followed by LDA to maximize class separability in the reduced feature space. The hybrid approach ensures that both global patterns and class-specific features are effectively preserved, improving classification performance. Experimental results on benchmark cancer microarray datasets demonstrate that the hybrid dimensionality reduction technique outperforms traditional methods, such as individual PCA and LDA, in terms of classification accuracy and computational efficiency. This method provides a promising solution to the challenges posed by high-dimensional genomic data, offering valuable insights for early cancer detection and personalized treatment strategies.