Michael Retchin edited Model Evaluation and Improvements.tex  over 9 years ago

Commit id: 00a410b3bd95cf019abe70aad270aca4141cc43e

deletions | additions      

       

\section{Model Evaluation and Improvements}  The model was evaluated for efficiency  by measuring the amount number  of generations. generations required to reach apparent-optima.  Based on this metric, the basic algorithm is efficient (i.e. not many generations are required to get a solution that is relatively seemingly  close to optimal). However, the model is not very accurate. A significant tradeoff between efficiency (the number of generations required to reach the optimal solution) and accuracy (F1 ($F_1$  score) was observed. To mitigate this problem, the model could be improved by introducing sexual reproduction into the genetic algorithm. Sexual reproduction would tend to find the optimal solution more efficiency because sexual reproduction produces greater genetic diversity than asexual reproduction.\cite{reed2003correlation}