Damn it CS you got me fall for the meme

classic

I feel you bro

advice: divide the parameters' intervals in 5 for each parameter, run the algorithm say 10 times for each combination of parameters, and evaluate some performance (speed of convergence, error or whatever).

do this automatically by using your already existing script in a quadruple loop, and don't forget to store the results.

This should give you a rough idea of what to do.

Let it run and go do something else.

>reading papers on anything
>if it used a survey/questionnaire they NEVER include it
>if they use a non-standard tool set-up, there's no pictures or in-depth description
>if it's electronic, they never include a circuit
>a study looking for correlations will check 10 parameters, find one correlation, but won't describe the correlation between the other elements
>no study will actually include the raw data, EVER, FUCKING EVER, IT'D BE A SMALLER SPREADSHEET THAN YOUR GODDAMN PDF YOU FUCKING AUTISTIC NERDS AAAAAAAAARGGGGGGGGGGHHHHHHHHHHHHH

THIS is how you fucking do a study (warning, not safe for electrical engineers):
http://www.eecs.berkeley.edu/Pubs/TechRpts/2011/EECS-2011-94.pdf

>>8108228
>use a non available function doesn't give any clue as to how it's evaluated

These niggas literally just gave it a name and kept using it while I was trying to reproduce their results.

>>8108228
A dissertation is not a paper. Nor is it a single study.

>>8107831
Maybe email one of the authors you stupid nigger

>>8108245
There's still no justification to why papers are half shit.

One of the dumbest examples: transcranial direct current stimulation. You take a voltage/current source and put it on the body. No complex circuitry needed. Literally some of the scientists use anode/cathode wrong, or at minimum, don't specify whether it's anode/cathode of the battery or stimulation site, which ended up with some of the papers hooking up electrodes backwards, and getting no or reverse results. One fucking picture would have resolved the issue.

Reminds me of a paper studying rats, the authors had to describe states of liking/disgust/whatever, so they put the fucking pictures in their report.

Don't get me started on that one Sociologist/Psychologist from Europe that had like ten years of work redacted because he was a fraud that made up the data and nobody realized like 50 people didn't exist and he never had to provide proof of it.

>>8108262
put drawings of the rats faces* to be clearer. One of the most obvious things for a study to do, yet it's not often people clarify what they mean when they say something like "rat facial expressions"

>>8108262
papers are limited to 8 pages (or sth like that) if you submit them for conferences.

Why do they include the lead author's email?

Dumb ass

>>8107831
Link the paper?

>>8107831
>> did manual parameters searching for 2 days

You could have, you know, emailed him.

>>8108228
>links to a tech report that appears as "Detailed Explanation can be found in [fuck]" in the actual paper
baka