Mathematical operations and functions
The datasets can be transposed and concatenated. If column names or row names in the concatenated datasets are not unique, this will be corrected by adding an additional letter to the names:
load people
d = people(1:4, 1:5);
show(d')
People:
People dataset
Objects
Lars Peter Rasmus Lene
----- ------ ------- -----
Height 198 184 183 166
Weight 92 84 83 47
Hairleng -1 -1 -1 -1
Shoesize 48 44 44 36
Age 48 33 37 32
show([d d])
Variables
Height Weight Hairleng Shoesize Age VHeight VWeight VHairleng VShoesize VAge
------- ------- --------- --------- ---- -------- -------- ---------- ---------- -----
Lars 198 92 -1 48 48 198 92 -1 48 48
Peter 184 84 -1 44 33 184 84 -1 44 33
Rasmus 183 83 -1 44 37 183 83 -1 44 37
Lene 166 47 -1 36 32 166 47 -1 36 32
show([d; d])
Variables
Height Weight Hairleng Shoesize Age
------- ------- --------- --------- ----
Lars 198 92 -1 48 48
Peter 184 84 -1 44 33
Rasmus 183 83 -1 44 37
Lene 166 47 -1 36 32
OLars 198 92 -1 48 48
OPeter 184 84 -1 44 33
ORasmus 183 83 -1 44 37
OLene 166 47 -1 36 32
The mdadata objects supports most of the basic arithmetic operations as well as some functions directly, converting all names automatically:
a = mdadata(reshape(1:15, 5, 3));
b = mdadata(reshape(15:-1:1, 5, 3));
show(a + b)
Variables
1 2 3
--- --- ---
16 16 16
16 16 16
16 16 16
16 16 16
16 16 16
show(a - b)
Variables
1 2 3
---- --- ---
-14 -4 6
-12 -2 8
-10 0 10
-8 2 12
-6 4 14
show(a * 5)
Variables
1 2 3
--- --- ---
5 30 55
10 35 60
15 40 65
20 45 70
25 50 75
show(a * b')
Variables
1 2 3 4 5
---- ---- ---- ---- ----
130 112 94 76 58
160 139 118 97 76
190 166 142 118 94
220 193 166 139 112
250 220 190 160 130
show(a .* b)
Variables
1 2 3
--- --- ---
15 60 55
28 63 48
39 64 39
48 63 28
55 60 15
show(a ./ b)
Variables
1 2 3
------- ------ -----
0.0667 0.6 2.2
0.143 0.778 3
0.231 1 4.33
0.333 1.29 7
0.455 1.67 15
Linear equations can be solved directly for mdadata objects.
load people
X = people(:, {'Weight', 'Beer', 'Swim'});
y = people(:, 'Height');
b = X\y;
show(b)
Height
--------
Weight -0.342
Beer -0.0147
Swim 2.44
show(X * b)
Height
-------
Lars 201
Peter 190
Rasmus 189
Lene 163
Mette 172
Gitte 173
Jens 174
Erik 172
Lotte 169
Heidi 168
Kaj 187
Gerda 151
Anne 164
Britta 162
Magnus 181
Casper 177
Luka 192
Federico 198
Dona 166
Fabrizia 164
Lisa 153
Benito 184
Franko 179
Alessandro 174
Leonora 164
Giuliana 161
Giovanni 174
Leonardo 169
Marta 166
Rosetta 165
Romeo 169
Romina 162
The full list of the functions includes: plus(), minus(), uminus(), times(), mtimes(), rdivide(), mrdivide(), ldivide(), mldivide(), round(), abs(), power(), sqrt(), log(), and exp().
a = mdadata(reshape(1:15, 5, 3));
show(log(a))
Variables
1 2 3
------ ----- -----
0 1.79 2.4
0.693 1.95 2.48
1.1 2.08 2.56
1.39 2.2 2.64
1.61 2.3 2.71
show(exp(a))
Variables
1 2 3
----- --------- ---------
2.72 403 5.99e+04
7.39 1.1e+03 1.63e+05
20.1 2.98e+03 4.42e+05
54.6 8.1e+03 1.2e+06
148 2.2e+04 3.27e+06
show(power(a, 0.5))
Variables
1 2 3
----- ----- -----
1 2.45 3.32
1.41 2.65 3.46
1.73 2.83 3.61
2 3 3.74
2.24 3.16 3.87
show(round(a/2))
Variables
1 2 3
-- -- --
1 3 6
1 4 6
2 4 7
2 5 7
3 5 8
Please, be aware that if a calculation results in a complex number, it will be converted to real and method will show a warning:
d = mdadata([-1 0.5; 1 0]);
show(log(d))
Warning: Complex values were obtained, converted to real.
Variables
1 2
-- -------
0 -0.693
0 -Inf
show(sqrt(d))
Warning: Complex values were obtained, converted to real.
Variables
1 2
-- ------
0 0.707
1 0
As an extra example we will calculate an BMI (Body Mass Index) of the persons in our People data and add it as an extra column to the original dataset. As one can see all names remain correct after the calculations.
load people
bmi = people(:, 'Weight') ./ (people(:, 'Height') / 100).^2;
bmi.colNames = {'BMI'};
newdata = [people bmi];
show(bmi(1:5));
BMI
-----
Lars 23.5
Peter 24.8
Rasmus 24.8
Lene 17.1
Mette 20.8
show(newdata(1:5, :));
Variables
Height Weight Hairleng Shoesize Age Income Beer Wine Sex Swim Region IQ BMI
------- ------- --------- --------- ---- -------- ----- ----- ---- ----- ------- ---- -----
Lars 198 92 -1 48 48 4.5e+04 420 115 -1 98 -1 100 23.5
Peter 184 84 -1 44 33 3.3e+04 350 102 -1 92 -1 130 24.8
Rasmus 183 83 -1 44 37 3.4e+04 320 98 -1 91 -1 127 24.8
Lene 166 47 -1 36 32 2.8e+04 270 78 1 75 -1 112 17.1
Mette 170 60 1 38 23 2e+04 312 99 1 81 -1 110 20.8
Calculation of size and length can be done using the same way as with matrices, plus there are two properties nRows and nCols. Function numel() was not overridden due to some technical reasons, so it will return 1 instead of number of values (use numel(d.values) if needed):
disp(people.nRows)
disp(people.nCols)
disp(size(people))
disp(length(people))
disp(numel(people))
32
12
32 12
32
1