Matlab - Modelformed.NumNeighbors

What should NumNeighbors be equal to so I can find the correct result? How is NumNeighbors determined?

I tried all the values, but some outputs show correct and then incorrect responses.

I have a Higher Education Students Performance Evaluation Dataset (145rows and 32columns)

opts = delimitedTextImportOptions("NumVariables", 32);
opts.DataLines = [1, Inf];
opts.Delimiter = ",";
opts.VariableNames = ["Q1", "Q2", "Q3", "Q4", "Q5", "Q6","Q7","Q8","Q9","Q10","Q11","Q12","Q13","Q14","Q15","Q16","Q17","Q18","Q19","Q20","Q21","Q22","Q23","Q24","Q25","Q26","Q27","Q28","Q29","Q30","Q31","Output"];
opts.VariableTypes = ["double", "double", "double", "double", "double","double", "double", "double", "double", "double", "double", "double", "double", "double", "double", "double", "double", "double", "double", "double", "double", "double", "double", "double", "double", "double", "double", "double", "double", "double", "double" "categorical"];
opts.ExtraColumnsRule = "ignore";
opts.EmptyLineRule = "read";
opts = setvaropts(opts, "Output", "EmptyFieldRule", "auto");
DATASET = readtable("HESPENUMERICDATAwithCategory.data", opts);
clear opts

modelformed = fitcknn(DATASET,'Output~Q1+Q2+Q3+Q4+Q5+Q6+Q7+Q8+Q9+Q10+Q11+Q12+Q13+Q14+Q15+Q16+Q17+Q18+Q19+Q20+Q21+Q22+Q23+Q24+Q25+Q26+Q27+Q28+Q29+Q30+Q31');
modelformed.NumNeighbors=31;
a1=input('Enter Student Age: ');
a2=input('Enter Sex: ');
a3=input('Enter Graduated high-school type: ');
a4=input('Enter Scholarship type: ');
a5=input('Enter Additional work: ');
a6=input('Enter Regular artistic or sports activity: ');
a7=input('Enter Do you have a partner: ');
a8=input('Enter Total salary if available: ');
a9=input('Enter Transportation to the university: ');
a10=input('Enter Accommodation type in Cyprus: ');
a11=input('Enter Mother education: ');
a12=input('Enter Father education: ');
a13=input('Enter Number of sisters/brothers (if available): ');
a14=input('Enter Parental status: ');
a15=input('Enter Mother occupation: ');
a16=input('Enter Father occupation: ');
a17=input('Enter Weekly study hours: ');
a18=input('Enter Reading frequency (non-scientific books/journals): ');
a19=input('Enter Reading frequency (scientific books/journals): ');
a20=input('Enter Attendance to the seminars/conferences related to the department: ');
a21=input('Enter Impact of your projects/activities on your success: ');
a22=input('Enter Attendance to classes: ');
a23=input('Enter Preparation to midterm exams 1: ');
a24=input('Enter Preparation to midterm exams 2: ');
a25=input('Enter Taking notes in classes: ');
a26=input('Enter Listening in classes: ');
a27=input('Enter Discussion improves my interest and success in the course: ');
a28=input('Enter Flip-classroom: ');
a29=input('Enter Cumulative grade point average in the last semester (/4.00): ');
a30=input('Enter Expected Cumulative grade point average in the graduation (/4.00): ');
a31=input('Enter Course ID: ');

predict(modelformed,[a1,a1,a3,a4,a5,a6,a7,a8,a9,a10,a11,a12,a13,a14,a15,a16,a17,a18,a19,a20,a21,a22,a23,a24,a25,a26,a27,a28,a29,a30,a31])

[Walter Roberson]

Are you asking what number of neighbors to use to get 100% prediction accuracy??

[Mustafa Furkan SAHIN]

Yes,that's exactly what i wanted to ask

[Walter Roberson]

As an outside observer (especially one who does not have access to your data), I have no reason to expect that getting 100% accuracy is even possible

[Mustafa Furkan SAHIN]

Thanks for your answer!

Answer 1

[Walter Roberson]

Your data has 145 rows. You can create a loop from 1 to (145-1)=144 neighbours, asking to predict with each of those number of neighbours in turn, and stopping as soon as you get 100% accuracy.

Hypothetically, accuracy might perhaps increase for a while as number of neighbours is increased, but then perhaps it would decrease. Or perhaps it would more or less reach a stable state, with more neighbours not making any difference. You should investigate.

[Mustafa Furkan SAHIN]

This answer is better, thanks a lot