This week I worked on my final report for DREU, performed my last pilot study for my project, and performed analysis on data from said pilot study with machine learning algorithms.
This day I worked on my final report for DREU.
Afterward, I attended the lab-wide "what are we going to do this week" meeting. I received the assignments of working on the data analysis for my study and to perform a pilot study on tuesday.
Afterward, I attended the weekly ITS meeting, where we discuss everything going on in the overarching ITS area. There I and the lab director worked out answers to those questions I had worked out with the grad student the previous week.
This day I ran what was the last pilot study for my project. All the flags were logged, the fNIRS data connection did not crash, everything went well.
Afterward, I attempted to run machine learning analysis on the data I collected from the pilot study. I could not get the program to successfully parse the files I was sending it, and scheduled to meet with a grad student in the lab to work out what was not working; in the meantime, they sent me some example files for me to assess what was wrong with mine.
This day I succeeded in running machine learning analysis on the data from the day previous's pilot study. I met with the grad student as above mentioned to discuss in what ways I should present the data to the machine learning analysis program and, afterward, spent some time getting said data into the decided-on configuration
The data comes in six channels, each of which has ~200 values per event and collectively has 31 events per channel. The decided-on data set consists of the minimum, maximum, average, median, and standard deviation for each event for each channel. Using the Attribute-Selected Classifier algorithm on this data set, its accuracy comes out at ~74%, which is higher than any other algorithm-data set pair tested.
This day I hit a wall with my final report and sent it to my grad student mentor to look at. I also reviewed the data analysis from the day before.
This day I attended the lab-wide weekly "what did you do this week?" meeting. I presented on having worked on my final report for DREU, on performing my last Pilot Study, and on the machine learning analysis I ran on the data from said pilot study.