Lung cancer diagnosis based on weighted convolutional neural network using gene data expression.

Area of Science:

• Oncology
• Bioinformatics
• Computational Biology

Background:

• Lung cancer is a significant global health burden, with increasing incidence and mortality rates.
• Existing prediction models face challenges like high dimensionality and computational inefficiency.
• Previous attribute selection methods and prediction algorithms have limitations in accuracy and scalability.

Purpose of the Study:

• To develop and evaluate a novel, efficient, and accurate technique for lung cancer prediction.
• To overcome the limitations of traditional methods such as multilayer perceptron and sequential minimal optimization (SMO).
• To enhance the precision, recall, and overall accuracy of lung cancer detection using a new computational approach.

Main Methods:

• Implementation of Z-score normalization for data preprocessing.
• Application of Levy Flight Cuckoo Search optimization for feature selection and parameter tuning.
• Development and utilization of a weighted convolutional neural network (CNN) for lung cancer prediction.
• Validation using the Kent Ridge Bio-Medical Dataset Repository.

Main Results:

• The proposed technique demonstrated superior performance in precision, recall, and accuracy compared to existing methods.
• The novel approach effectively addressed the computational inefficiencies and parameter challenges of previous models.
• Successful prediction of lung cancer using the integrated Z-score normalization, Levy Flight Cuckoo Search, and weighted CNN.

Conclusions:

• The novel technique offers a more effective and efficient solution for lung cancer prediction.
• This approach holds promise for improving early detection and patient outcomes in oncology.
• The study highlights the potential of combining advanced optimization and deep learning for complex disease prediction.