Ability to estimate cycle time is becoming more important for organisations in order to better utilize resources and potentially save money. The aim is to research whether three different machine learning techniques can be used to do this estimation of cycle time in a food processing environment and which one gets the highest accuracy.
Accuracy is tested with the help of three error measurements, called Mean Absolute Error, Root Mean Squared Error and Mean Absolute Percentage Error, for each technique. The implementation of Artificial Neural Network, Decision Tree and k-Nearest Neighbor were usedwith a Python library called scikit-learn. Data from a real slaughterhouse was provided and used to train and test the techniques.
Results shows that all techniques managed to produce relatively equal estimations for this problem. The better technique ended up being decision tree reached a Mean Absolute Percentage Error of 34.77%, closely followed by KNN with 35.06%.