This study aims to create new methods from problems that we will meet or face in ICIS 2018 Challgne that commen method was not the best option for these problems
Results are 2 methods one in segmentation called limited crop certain(LCC) its pseudocode (to fully understand it or to know how to use it read document by recommendation)
and one in enseble called class weight transformation(CWT) its psudocode (to fully understand it or to know how to use it read document by recommendation)
we did not aim to get high results we just want to prove that our methods worth to try as replacement for common methods in this field
heightest results we could achive was by novel enseble (CWT) (best 10 simple models no 5-fold) 83,4% balanced accurcy on live leader board without extra data within our experiments
Best first teams of this challenge
(LCC) was compared with crop certain(crop based on prediction of musk) by simple analysis and with bitwise(AND mask with image) by models created to compare limited crop to bitwise.
(CWT) was compared with stacking ensemble (average and voting)
Important note : all notebooks were created on Google colab due our limited resources so path system would be different if you used it on your laptop and you will be forced to download all data from drivers instead of just including it as add to my drive from links that we provide in some of the notebooks.
recommendation : if you want to understand just the 2 novel method and not interested in anything else foucs reading only on start of explaination of each method in both chapters('page:' pre-processing chapter for (LCC) ,'page:' training and test results chapter for (CWT)) .
Thanks for being patient ,we hope you enjoy reading the 2 novel methods
