Implementation of New Efficient Method to Analyze Texture of Plant Based Alternative Meat Using Machine Learning Techniques

The firming on meat production are not capable to produce sufficient meat to current market due to animal safety and welfare. General problems faced when culture meat are many people are not interest on firming, space problem to grow animals, and animals are frequently effected various types of diseases. All these factors forces a challenging task to produce required quantity of meat to market that will provide a chance to making alternative meat. Plant based meat is a solution to this problem and alternative meat must appear with same texture, crunchiness, spicy, taste like real meat. The manufacturing process of plant based meat required more number of repeated experiments with expert domain knowledge to meet consumer requirements. To provide solutions for these problems we developed a new model to analyze food texture using machine learning methods. We used different types of ingredients like fat, fish, fiber, protein, moisture to meet the taste of natural meet, and all these values are taken as a input to ML methods. In our research we used ML methods naive Bayes, random forest, XGBoost, KNN, and Decision Tree. With help of these methods features like chewiness and hardness of alternative meat are measured. 19% MAPE and 8.9% RMSE values are estimated under hardness category, 12% MAPE and 5.98% RMSE values are estimated under chewiness.