Skip to content

Commit

Permalink
fixed coding style, Fixed docstring, added Demo and BISTs
Browse files Browse the repository at this point in the history
  • Loading branch information
@AZM999
AZM999 committed Jul 16, 2023
1 parent a41cc29 commit 2a528ed
Showing 1 changed file with 28 additions and 27 deletions.
55 changes: 28 additions & 27 deletions inst/fitcknn.m
View file
Original file line number Diff line number Diff line change
Expand Up @@ -47,12 +47,12 @@
## in the kNN search. It must be a positive integer value and by default it is
## 1. @var{"K"} can be changed by using @code{@var{obj.K} = 10}.
##
## @item @tab @qcode{"weights"} @tab is a @math{Nx1} numeric non-negative matrix of
## the observational weights, each row in @var{weights} corresponds to the row
## in @var{Y} and indicates the relative importance or weight to be considered
## in calculating the Nearest-neighbour, negative values are removed before
## calculations if weights are specified. this propery cannot be changed via
## object. default value @qcode{@var{weight} = ones(rows(Y),1)}.
## @item @tab @qcode{"weights"} @tab is a @math{Nx1} numeric non-negative matrix
## of the observational weights, each row in @var{weights} corresponds to the
## row in @var{Y} and indicates the relative importance or weight to be
## considered in calculating the Nearest-neighbour, negative values are removed
## before calculations if weights are specified. this propery cannot be
## changed via object. default value @qcode{@var{weight} = ones(rows(Y),1)}.
##
## @item @tab @qcode{"cost"} @tab is a @math{NxR} numeric matrix containing
## misclassification cost for the corresponding instances in @var{X} where
Expand Down Expand Up @@ -185,12 +185,12 @@
## create object
if ( nargin == 0)
## return empty object with warning
obj = classificationKNN();
obj = ClassificationKNN();
warning ("fitcknn: No Argument given, Created Object will be empty.");

else
## arguments are not empty and are within range
obj = classificationKNN(X, Y, varargin{:});
obj = ClassificationKNN(X, Y, varargin{:});
endif


Expand All @@ -209,28 +209,29 @@
%! ## create an object
%! a = fitcknn (x, y, 'Xclass' , xnew, 'k', 5)
%! ## predict labels for points in xnew
%! a.predict()
%! predict (a)
%! ## change properties keeping training data and predict again
%! a.distance = 'hamming';
%! a.k = 10;
%! a.predict()
%! predict (a)

## Test Output
%!test
%! warning("off");
%! a = fitcknn ();
%! assert (class (a), "classificationKNN");
%! assert (class (a), "ClassificationKNN");
%! assert ({a.Breakties, a.Includeties, a.NN, a.NSmethod}, {[],[],[],[]})
%! assert ({a.NosClasses, a.NumObsv, a.Scale, a.Score, a.cost}, {[],[],[],[],[]})
%! assert ({a.X, a.Xclass, a.Y, a.bucketsize, a.classNames}, {[],[],[],[],[]})
%! assert ({a.cov, a.distance, a.exponent, a.k, a.label}, {[],[],[],[],[]})
%! assert ({a.prior, a.standerdize, a.weights}, {[],[],[]})
%! assert ({a.cov, a.distance, a.P, a.k, a.label}, {[],[],[],[],[]})
%! assert ({a.prior, a.standardize, a.weights}, {[],[],[]})

%!test
%! warning("off");
%! x = [1,2,3;4,5,6;7,8,9;3,2,1];
%! y = ['a';'a';'b';'b'];
%! a = fitcknn (x, y);
%! assert (class (a), "classificationKNN");
%! assert (class (a), "ClassificationKNN");
%! assert ({a.X, a.Y, a.k},{[1,2,3;4,5,6;7,8,9;3,2,1], ['a';'a';'b';'b'], 1})
%! assert ({a.NSmethod, a.distance, a.bucketsize},{"exhaustive","euclidean",50})

Expand All @@ -240,7 +241,7 @@
%! y = ['a';'a';'b';'b'];
%! k = 10;
%! a = fitcknn (x, y, "K" ,k);
%! assert (class (a), "classificationKNN");
%! assert (class (a), "ClassificationKNN");
%! assert ({a.X, a.Y, a.k},{[1,2,3;4,5,6;7,8,9;3,2,1], ['a';'a';'b';'b'], 10})
%! assert ({a.NSmethod, a.distance, a.bucketsize},{"exhaustive","euclidean",50})

Expand All @@ -250,7 +251,7 @@
%! y = ['a';'a';'b';'b'];
%! weights = ones (4,1);
%! a = fitcknn (x, y, "weights" , weights);
%! assert (class (a), "classificationKNN");
%! assert (class (a), "ClassificationKNN");
%! assert ({a.X, a.Y, a.k},{[1,2,3;4,5,6;7,8,9;3,2,1], ['a';'a';'b';'b'], 1})
%! assert (a.weights, ones (4,1))
%! assert ({a.NSmethod, a.distance, a.bucketsize},{"exhaustive","euclidean",50})
Expand All @@ -259,10 +260,10 @@
%! warning("off");
%! x = [1,2,3;4,5,6;7,8,9;3,2,1];
%! y = ['a';'a';'b';'b'];
%! a = fitcknn (x, y, "exponent" , 10);
%! assert (class (a), "classificationKNN");
%! a = fitcknn (x, y, "p" , 10);
%! assert (class (a), "ClassificationKNN");
%! assert ({a.X, a.Y, a.k},{[1,2,3;4,5,6;7,8,9;3,2,1], ['a';'a';'b';'b'], 1})
%! assert (a.exponent, 10)
%! assert (a.P, 10)
%! assert ({a.NSmethod, a.distance, a.bucketsize},{"exhaustive","euclidean",50})

%!test
Expand All @@ -271,7 +272,7 @@
%! y = ['a';'a';'b';'b'];
%! cov = rand (4,1);
%! a = fitcknn (x, y, "cov" , cov, 'distance', 'mahalanobis');
%! assert (class (a), "classificationKNN");
%! assert (class (a), "ClassificationKNN");
%! assert ({a.X, a.Y, a.k},{[1,2,3;4,5,6;7,8,9;3,2,1], ['a';'a';'b';'b'], 1})
%! assert (a.cov, cov)
%! assert ({a.NSmethod, a.distance, a.bucketsize},{"exhaustive","mahalanobis",50})
Expand All @@ -281,26 +282,26 @@
%! x = [1,2,3;4,5,6;7,8,9;3,2,1];
%! y = ['a';'a';'b';'b'];
%! a = fitcknn (x, y, "bucketsize" , 20, 'distance', 'mahalanobis');
%! assert (class (a), "classificationKNN");
%! assert (class (a), "ClassificationKNN");
%! assert ({a.X, a.Y, a.k},{[1,2,3;4,5,6;7,8,9;3,2,1], ['a';'a';'b';'b'], 1})
%! assert ({a.NSmethod,a.distance,a.bucketsize},{"exhaustive","mahalanobis",20})

%!test
%! warning("off");
%! x = [1,2,3;4,5,6;7,8,9;3,2,1];
%! y = ['a';'a';'b';'b'];
%! a = fitcknn (x, y, 'standerdize', true);
%! assert (class (a), "classificationKNN");
%! a = fitcknn (x, y, 'standardize', true);
%! assert (class (a), "ClassificationKNN");
%! assert ({a.X, a.Y, a.k},{[1,2,3;4,5,6;7,8,9;3,2,1], ['a';'a';'b';'b'], 1})
%! assert (a.standerdize, true);
%! assert (a.standardize, true);
%! assert ({a.NSmethod,a.distance,a.bucketsize},{"exhaustive","euclidean",50})

%!test
%! warning("off");
%! x = [1,2,3;4,5,6;7,8,9;3,2,1];
%! y = ['a';'a';'b';'b'];
%! a = fitcknn (x, y, 'includeties', true);
%! assert (class (a), "classificationKNN");
%! assert (class (a), "ClassificationKNN");
%! assert ({a.X, a.Y, a.k},{[1,2,3;4,5,6;7,8,9;3,2,1], ['a';'a';'b';'b'], 1})
%! assert (a.Includeties, true);
%! assert ({a.NSmethod,a.distance,a.bucketsize},{"exhaustive","euclidean",50})
Expand All @@ -311,7 +312,7 @@
%! y = ['a';'a';'b';'b'];
%! cost = ones (4,2);
%! a = fitcknn (x, y, 'cost', cost );
%! assert (class (a), "classificationKNN")
%! assert (class (a), "ClassificationKNN")
%! assert ({a.X, a.Y, a.k},{[1,2,3;4,5,6;7,8,9;3,2,1], ['a';'a';'b';'b'], 1})
%! assert (a.cost, [1,1;1,1;1,1;1,1])
%! assert ({a.NSmethod,a.distance,a.bucketsize},{"exhaustive","euclidean",50})
Expand All @@ -322,7 +323,7 @@
%! y = ['a';'a';'b';'b'];
%! scale = [1,2,3,4];
%! a = fitcknn (x, y, 'scale', scale );
%! assert (class (a), "classificationKNN")
%! assert (class (a), "ClassificationKNN")
%! assert ({a.X, a.Y, a.k},{[1,2,3;4,5,6;7,8,9;3,2,1], ['a';'a';'b';'b'], 1})
%! assert (a.Scale, [1,2,3,4])
%! assert ({a.NSmethod,a.distance,a.bucketsize},{"exhaustive","euclidean",50})
Expand Down

0 comments on commit 2a528ed

Please sign in to comment.