An Approach to avoid SSS problem in the data space using Fisher Face(PCA+LDA ) technique: A Case Study on Chest X-ray Pneumonia data

• Published in: 2024 IEEE International Students' Conference on Electrical, Electronics and Computer Science (SCEECS) pp. 1 - 6
• Main Authors: Rao, D. Rajeswara, Pathakamudi, Bhogeswar, Dakshayani, Reddi
• Format: Conference Proceeding
• Language: English
• Published: IEEE 24.02.2024
• Subjects: Chest X-ray of Pneumonia dataset; Feature extraction; Forestry; Linear discriminant analysis; Noise; PCA+LDA technique(Fisher Face technique); Pneumonia; Random Forest Algorithm; Reliability; SSS problem; X-ray imaging
• ISSN: 2688-0288
• DOI: 10.1109/SCEECS61402.2024.10482363

Pneumonia is a complex disease with a multitude of factors influencing its development and progression. Analysing Pneumonia data can be challenging due to the high dimensionality and noise present in the data. This study presents a methodology for identifying hidden patterns in chest X-ray images of pneumonia patients using Fisher face technique, which uses both principal component analysis (PCA) and linear discriminant analysis (LDA) i.e pca+lda technique (Fisher face technique), as well as the Random Forest algorithm. The SSS (small sample size) problem is a well-known issue that can arise when using LDA (Linear Discriminant Analysis) for data analysis. This problem occurs when the number of variables (features) in the data is larger than the number of samples, leading to a situation where the covariance matrix cannot be estimated reliably PCA is used to reduce the dimensionality of the chest X-ray images by identifying the most relevant features, while LDA is used to further enhance the separation between classes. The Random Forest classification algorithm is then used to classify the chest X-ray images based on the extracted features.