yilmazduygu / matlab-final Goto Github PK

• Dont ignore Matlab's warnings - you should end up with a green light in the top right corner of the editor.

• cell2mat or cat - I think either one is fine. With cat (and using dim 1) it may be possible to get rid of the somewhat klunky transpose you now need on the A'

• datetime always stores a specific time point (so always a date and a time). If you don't specify a date it will assume the time refers to today. I would fix this by storing the date as a char in a local variable (recDate) in the 'start date' section and then using this recDate together with the time in the 'start time' section to create a single datetime to be stored in the S struct.

• Reduce duplication. Write a function that reads one of those data blocks (e.g. C or D) and call this function three times from your import script/function. As this function is unlikely to be useful outside the context of this specific import function, it could be defined inside the same function m-file.

• The structure of the import routine is fine, but needs some error checking. For instance in your comments you write " Now fid should be right below the line "A: "" - you should check that this is the case and generate an error if it isnt'. The same goes for the if line(1)=='C" and later statements, what if the line(1) is not 'C'?

• because you are not using the tag when reading the C and D parts, you could use extractAfter instead of split (and you could use str2num instead of sscanf). With those changes you could replace l175: 179 with a single line.

Hi Bart and Katie!
I have some issues with the analysis part;

• I cannot figure out how to label the outliers programmatically.

• My data is only the training data (I was hoping to get the test data as well by now, but unfortunately I couldn't), that's why I am doing more basic descriptive analyses. Is that ok?

Thanks!

The base code looks fine, but you could improve the modularity, and the structure quite a bit and make this code more reusable.

• Separate functions for the conversion from the Medx files - a function that takes a file name or a path for instance, and a destination path for the analysis.mat

• Separate functions for graphics (that take the analysis.mat file as its input) - that way you can analyze different subsets

• Your functions need help sections.