Advanced Technique for Purifying Chitin Prediction from Crustacean Biowaste Utilizing Dynamic Path Controllable Deep Unfolding Networks for Enhanced Purity Improvement

Abstract

Chitin, a natural polymer found in crab and mollusk shells, insects and fungi. Due to its biological applications, chitin has been recovered as hydrogel, nanoparticles, nanosheets, and nanowires. Deep eutectic solvents are gaining attention for being environmentally beneficial and recyclable. To obtain pure chitin with higher yield, many factors must be considered. It's crucial and nevertheless difficult to create novel methods for making high-quality chitin. In this manuscript, Advanced Technique for Purifying Chitin Prediction from Crustacean Biowaste Utilizing Dynamic Path Controllable Deep Unfolding Networks for Enhanced Purity Improvement (PC-CB-DPCDUN-PI) is proposed. Here, the data is taken from Chitin Decomposition 16S rRNA gene fastq dataset. The gathered data is preprocessing and normalized using Generalized Multi-kernel Maximum Correntropy Kalman Filter (GMMCKF). The pre-processed data is fed into the Dynamic Path Controllable Deep Unfolding Networks (DPCDUN) to predict chitin purity from various crustacean shell waste using deep eutectic solvents (DES). Finally, Snow Avalanches Algorithm (SAA) is proposed to optimize the weight parameters of DPCDUN for effective chitin purity prediction. The PC-CB-DPCDUN-PI method demonstrates better performance by achieving 6.68%, 15.75%, and 10.16% lower RMSE and 10.03%, 12.11% and 8.62% higher R² compared to the existing CCS-CSP-ANN, AVP-SSW-GP, and FVW-CAP-TGA methods respectively, confirming its effectiveness in enhancing chitin purity prediction from crustacean bio-waste.