Crab Age Prediction using Machine Learning for Sustainable Crab Farming

Seafood has played an important role in the diet of quite a large number of people in the world, in particular crabs are considered nutritious and delicious in many different regions and countries. Consequently, crab farms are profitable because of the demand. Managing the populations and harvesting are an important indicator of the crab farms profitability. Therefore, predicting the age of crabs can be quite beneficial. However, using traditional methods predicting age can be challenging. In this paper, we proposed a method to predict the age of crabs using machine learning techniques. We explored the use of several regression models like xgboost, random forest, linear regression and elastic net regression. Extensive experimentation using holdout and k-fold cross validation indicated that the linear regression model outperformed the other models. The linear regression model achieved a MAE of 1.36, and a MAPE of 14.6. This method provides an autonomous and a more efficient method of predicting ages which can assist crab farmers in crucial decisions. This will not only benefit the business in terms of costs and resource management, but most importantly, it will help the environment by introducing more sustainable farming practices and putting less pressure on the local ecosystems by controlling the population.