###### tags: `Meeting` # Meeting Sumary of 20.06.2021 * To get the answers to my questions about base concepts; I need to read the SMOTEBoost paper with implementation details and also review the AdaBoost and AdaFair paper once more (now with more details). I will summarize the detals of differences between these boosting methods in a file and present it for you on 25.06.2021 (our next meeting). * To find out about parameters and how to tune them I will understand them in the code and in the next meeting breifly discuss them again. * n_sample is percentage of what ? protected group or percentage of which data it is... * Next step in implementation is to detect how the synthetic data contribute to different groups by visualizing the instances: * Have a visual analysis of data (scatters including each group) to see how the synthetic data is added to each group. * I'll search to find a descriptive plot. * Send an example of the plot to you before our next meeting on 25.06 and ask for your acknowledgment.