1 /* Example inspired by the toy example in https://www.r-bloggers.com/2020/06/understanding-lasso-and-ridge-regression-2/
2 * blog post by Dr. Atakan Ekiz.
3 * Here we wish to predict the number of shark attacks (that is, this number is our response variable),
4 * using the following predictor variables:
5 * - percentage of swimmers who watched the movie Jaws
6 * - the number of swimmers in the water
7 * - the average temperature of the day
8 * - the price of your favorite tech stock of the day (totally uncorrelated variable) */
9
10 static char help[] = "Tests basic creation and destruction of PetscRegressor objects.\n\n";
11
12 #include <petscregressor.h>
13
main(int argc,char ** args)14 int main(int argc, char **args)
15 {
16 PetscRegressor regressor;
17 PetscMPIInt rank;
18 Mat X;
19 Vec y, y_predicted, coefficients;
20 PetscScalar intercept;
21
22 PetscScalar y_array[20] = {98, 53, 39, 127, 73, 42, 71, 61, 83, 74, 85, 82, 62, 60, 43, 69, 67, 69, 85, 3}; // Number of shark attacks
23
24 PetscScalar X_array[80] = {37.92934, 513, 92.89899, 137.2139, // % watched Jaws, #swimmers, temperature, stock price
25 52.77429, 451, 87.86271, 145.7987, //
26 60.84441, 456, 88.28927, 149.7299, //
27 26.54302, 546, 89.43875, 147.1180, //
28 54.29125, 431, 88.01132, 124.3068, //
29 55.06056, 355, 88.06297, 114.1730, //
30 44.25260, 557, 87.78536, 112.5773, //
31 44.53368, 398, 87.49603, 125.1628, //
32 44.35548, 498, 88.95234, 124.8483, //
33 41.09962, 406, 89.00630, 115.9223, //
34 45.22807, 610, 86.38794, 148.1111, //
35 40.01614, 452, 88.83585, 131.7050, //
36 42.23746, 429, 87.78222, 106.3717, //
37 50.64459, 450, 87.97008, 121.1523, //
38 59.59494, 337, 89.67538, 145.7158, //
39 48.89715, 383, 91.12611, 123.3896, //
40 44.88990, 282, 93.29563, 145.4085, //
41 40.88805, 366, 88.45329, 129.8872, //
42 41.62828, 471, 93.21182, 131.5871, //
43 74.15835, 453, 87.68438, 143.4579};
44
45 PetscInt rows_ix[20] = {0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19};
46 PetscInt cols_ix[4] = {0, 1, 2, 3};
47
48 PetscCall(PetscInitialize(&argc, &args, (char *)0, help));
49 PetscCallMPI(MPI_Comm_rank(PETSC_COMM_WORLD, &rank));
50
51 PetscCall(VecCreate(PETSC_COMM_WORLD, &y));
52 PetscCall(VecSetSizes(y, PETSC_DECIDE, 20));
53 PetscCall(VecSetFromOptions(y));
54 PetscCall(VecDuplicate(y, &y_predicted));
55 PetscCall(MatCreate(PETSC_COMM_WORLD, &X));
56 PetscCall(MatSetSizes(X, PETSC_DECIDE, PETSC_DECIDE, 20, 4));
57 PetscCall(MatSetFromOptions(X));
58 PetscCall(MatSetUp(X));
59
60 if (!rank) {
61 PetscCall(VecSetValues(y, 20, rows_ix, y_array, INSERT_VALUES));
62 PetscCall(MatSetValues(X, 20, rows_ix, 4, cols_ix, X_array, ADD_VALUES));
63 }
64 PetscCall(VecAssemblyBegin(y));
65 PetscCall(VecAssemblyEnd(y));
66 PetscCall(MatAssemblyBegin(X, MAT_FINAL_ASSEMBLY));
67 PetscCall(MatAssemblyEnd(X, MAT_FINAL_ASSEMBLY));
68
69 PetscCall(PetscRegressorCreate(PETSC_COMM_WORLD, ®ressor));
70 PetscCall(PetscRegressorSetType(regressor, PETSCREGRESSORLINEAR));
71 PetscRegressorSetFromOptions(regressor);
72 PetscCall(PetscRegressorFit(regressor, X, y));
73 PetscCall(PetscRegressorPredict(regressor, X, y_predicted));
74 PetscCall(PetscRegressorLinearGetIntercept(regressor, &intercept));
75 PetscCall(PetscRegressorLinearGetCoefficients(regressor, &coefficients));
76
77 PetscCall(PetscPrintf(PETSC_COMM_WORLD, "Intercept is %lf\n", intercept));
78 PetscCall(PetscPrintf(PETSC_COMM_WORLD, "Coefficients are\n"));
79 PetscCall(VecView(coefficients, PETSC_VIEWER_STDOUT_WORLD));
80 PetscCall(PetscPrintf(PETSC_COMM_WORLD, "Predicted values are\n"));
81 PetscCall(VecView(y_predicted, PETSC_VIEWER_STDOUT_WORLD));
82
83 PetscCall(PetscRegressorDestroy(®ressor));
84
85 PetscCall(PetscFinalize());
86 return 0;
87 }
88