#include <stdlib.h>
#include <stdio.h>
#include <math.h>

#define N_FEATURES 6
#define N_CLASSES 5
#define N_VECTORS 57
#define N_ROWS 5
#define N_COEFFICIENTS 4
#define N_INTERCEPTS 10
#define KERNEL_TYPE 'l'
#define KERNEL_GAMMA 0.02
#define KERNEL_COEF 0.0
#define KERNEL_DEGREE 3

double vectors[57][6] = {{19.0, 67.0, 203.0, 193.0, 129.0, 148.0}, {19.0, 81.0, 217.0, 200.0, 131.0, 153.0}, {19.0, 77.0, 211.0, 197.0, 130.0, 151.0}, {18.0, 74.0, 216.0, 200.0, 131.0, 150.0}, {8.0, 83.0, 220.0, 187.0, 144.0, 145.0}, {7.0, 106.0, 163.0, 132.0, 146.0, 145.0}, {7.0, 129.0, 182.0, 139.0, 152.0, 152.0}, {7.0, 128.0, 182.0, 139.0, 152.0, 152.0}, {9.0, 93.0, 245.0, 204.0, 146.0, 151.0}, {2.0, 118.0, 206.0, 152.0, 160.0, 147.0}, {2.0, 123.0, 224.0, 162.0, 164.0, 150.0}, {1.0, 122.0, 157.0, 114.0, 157.0, 143.0}, {4.0, 132.0, 178.0, 130.0, 157.0, 150.0}, {1.0, 153.0, 152.0, 100.0, 164.0, 148.0}, {1.0, 150.0, 173.0, 115.0, 167.0, 150.0}, {2.0, 151.0, 152.0, 102.0, 162.0, 149.0}, {1.0, 159.0, 162.0, 104.0, 167.0, 151.0}, {178.0, 124.0, 169.0, 119.0, 162.0, 140.0}, {177.0, 152.0, 121.0, 78.0, 160.0, 140.0}, {178.0, 155.0, 115.0, 73.0, 159.0, 141.0}, {144.0, 169.0, 124.0, 76.0, 163.0, 146.0}, {1.0, 165.0, 150.0, 94.0, 167.0, 150.0}, {178.0, 149.0, 119.0, 78.0, 159.0, 141.0}, {164.0, 155.0, 119.0, 76.0, 160.0, 142.0}, {1.0, 136.0, 156.0, 108.0, 160.0, 145.0}, {178.0, 145.0, 128.0, 85.0, 160.0, 140.0}, {114.0, 157.0, 119.0, 76.0, 160.0, 143.0}, {168.0, 150.0, 115.0, 74.0, 158.0, 141.0}, {0.0, 159.0, 157.0, 100.0, 167.0, 150.0}, {177.0, 160.0, 118.0, 74.0, 161.0, 141.0}, {0.0, 133.0, 191.0, 131.0, 167.0, 147.0}, {178.0, 152.0, 124.0, 80.0, 160.0, 141.0}, {178.0, 122.0, 162.0, 116.0, 160.0, 139.0}, {178.0, 157.0, 127.0, 81.0, 162.0, 142.0}, {178.0, 147.0, 122.0, 80.0, 159.0, 140.0}, {1.0, 135.0, 149.0, 104.0, 159.0, 145.0}, {0.0, 138.0, 198.0, 133.0, 169.0, 149.0}, {179.0, 125.0, 167.0, 117.0, 161.0, 142.0}, {178.0, 157.0, 114.0, 72.0, 159.0, 141.0}, {1.0, 129.0, 182.0, 127.0, 164.0, 146.0}, {1.0, 132.0, 159.0, 112.0, 160.0, 145.0}, {178.0, 151.0, 116.0, 75.0, 158.0, 140.0}, {177.0, 173.0, 122.0, 73.0, 164.0, 144.0}, {177.0, 165.0, 128.0, 79.0, 164.0, 143.0}, {178.0, 166.0, 122.0, 75.0, 163.0, 143.0}, {177.0, 118.0, 120.0, 87.0, 153.0, 136.0}, {178.0, 115.0, 135.0, 99.0, 154.0, 137.0}, {177.0, 118.0, 122.0, 88.0, 153.0, 136.0}, {124.0, 114.0, 123.0, 91.0, 151.0, 137.0}, {177.0, 171.0, 114.0, 69.0, 162.0, 143.0}, {177.0, 172.0, 113.0, 68.0, 162.0, 143.0}, {176.0, 144.0, 123.0, 81.0, 159.0, 138.0}, {178.0, 167.0, 118.0, 72.0, 162.0, 143.0}, {177.0, 172.0, 123.0, 75.0, 164.0, 144.0}, {177.0, 135.0, 134.0, 92.0, 159.0, 138.0}, {177.0, 120.0, 139.0, 100.0, 156.0, 137.0}, {178.0, 167.0, 119.0, 73.0, 162.0, 143.0}};
double coefficients[4][57] = {{0.0, 0.0038545776095045424, 0.0, 5.3340301342833886e-05, -0.0033247158696419685, -0.0, -0.0, -0.0, -0.0005832020412054078, -7.057383918306568e-05, -0.00036143565337392973, -0.0, -0.0, -0.0, -0.0, -0.0, -0.0, -1.84284223988671e-05, -0.0, -0.0, -0.0, -0.0, -0.0, -0.0, -0.0, -0.0, -0.0, -0.0, -0.0, -0.0, -0.00019441952851682932, -0.0, -0.0, -0.0, -0.0, -0.0, -0.0, -0.0, -0.0, -0.0, -0.0, -0.0, -0.0, -0.0, -0.0, -0.0, -0.0, -0.0, -6.526776098945923e-05, -0.0, -0.0, -0.0, -0.0, -0.0, -0.0, -0.0, -0.0}, {0.0, 0.0004320094925569954, 0.0, 0.0, 0.0, 0.0, 0.025302498031903337, 0.00211378176605504, 0.0, -0.006016072596945847, -0.0, -0.0, -0.021400207201012534, -0.0, -0.0, -0.0, -0.0, -0.0, -0.0, -0.0, -0.0, -0.0, -0.0, -0.0, -0.0, -0.0, -0.0, -0.0, -0.0, -0.0, -0.0006153624681027348, -0.0, -0.0, -0.0, -0.0, -0.0, -0.0, -0.00019438819844166922, -0.0, -0.005318489351289106, -0.0, -0.0, -0.0, -0.0, -0.0, -0.0, -0.0, -0.0, -0.00011680196227296622, -0.0, -0.0, -0.0, -0.0, -0.0, -0.0, -0.0, -0.0}, {0.0, 0.0, 0.00021284795091569642, 0.0, 0.0, 0.00022676527329307724, 0.005901474744540431, 0.0, 0.0, 0.0, 0.0, 1.0, 0.0, 1.0, 0.03514172035883584, 0.0, 1.0, -0.0, -0.0, -0.0, -0.002110636488197338, -1.0, -0.0, -0.0, -0.09241440956197279, -0.0, -0.0, -0.0, -0.05274462280859587, -0.0, -0.0, -0.0, -0.0, -0.0, -0.0, -0.6213567548696584, -0.2665152966304088, -0.0, -0.0, -0.0, -1.0, -0.0, -0.0, -0.0, -0.0, -0.0, -0.0, -0.0, -0.00011847542874431661, -0.0, -0.0, -0.0, -0.0, -0.0, -0.0, -0.0, -0.0}, {6.526776098945923e-05, 0.0, 0.0, 0.0, 0.0, 0.00011680196227296622, 0.0, 0.0, 0.0, 0.0, 0.0, 8.837019916436589e-05, 0.0, 0.0, 0.0, 3.0105229579950734e-05, 0.0, 0.0, 1.0, 1.0, 0.0, 0.0, 1.0, 1.0, 0.0, 1.0, 0.6196007776877644, 0.014877240287376452, 0.0, 1.0, 0.0, 1.0, 1.0, 1.0, 1.0, 0.0, 0.0, 0.0, 1.0, 0.0, 0.0, 1.0, -1.0, -1.0, -1.0, -0.2525249717049514, -0.9139145934471427, -1.0, -0.8694337173117702, -1.0, -0.3487089413240337, -1.0, -1.0, -1.0, -1.0, -0.2498957941872232, -1.0}};
double intercepts[10] = {1.4580565469820796, 0.857852182962587, -1.0407286545400531, -0.9300982279802896, 7.657342570536519, -0.5929370566689991, -0.24919698650937327, -42.8270886357245, -0.3486237873011033, -113.99951084737917};
int weights[5] = {4, 5, 8, 25, 15};

int predict (double features[]) {
    int i, j, k, d, l;

    double kernels[N_VECTORS];
    double kernel;
    switch (KERNEL_TYPE) {
        case 'l':
            // <x,x'>
            for (i = 0; i < N_VECTORS; i++) {
                kernel = 0.;
                for (j = 0; j < N_FEATURES; j++) {
                    kernel += vectors[i][j] * features[j];
                }
                kernels[i] = kernel;
            }
            break;
        case 'p':
            // (y<x,x'>+r)^d
            for (i = 0; i < N_VECTORS; i++) {
                kernel = 0.;
                for (j = 0; j < N_FEATURES; j++) {
                    kernel += vectors[i][j] * features[j];
                }
                kernels[i] = pow((KERNEL_GAMMA * kernel) + KERNEL_COEF, KERNEL_DEGREE);
            }
            break;
        case 'r':
            // exp(-y|x-x'|^2)
            for (i = 0; i < N_VECTORS; i++) {
                kernel = 0.;
                for (j = 0; j < N_FEATURES; j++) {
                    kernel += pow(vectors[i][j] - features[j], 2);
                }
                kernels[i] = exp(-KERNEL_GAMMA * kernel);
            }
            break;
        case 's':
            // tanh(y<x,x'>+r)
            for (i = 0; i < N_VECTORS; i++) {
                kernel = 0.;
                for (j = 0; j < N_FEATURES; j++) {
                    kernel += vectors[i][j] * features[j];
                }
                kernels[i] = tanh((KERNEL_GAMMA * kernel) + KERNEL_COEF);
            }
            break;
    }

    int starts[N_ROWS];
    int start;
    for (i = 0; i < N_ROWS; i++) {
        if (i != 0) {
            start = 0;
            for (j = 0; j < i; j++) {
                start += weights[j];
            }
            starts[i] = start;
        } else {
            starts[0] = 0;
        }
    }

    int ends[N_ROWS];
    for (i = 0; i < N_ROWS; i++) {
        ends[i] = weights[i] + starts[i];
    }

    if (N_CLASSES == 2) {

        for (i = 0; i < N_VECTORS; i++) {
            kernels[i] = -kernels[i];
        }

        double decision = 0.;
        for (k = starts[1]; k < ends[1]; k++) {
            decision += kernels[k] * coefficients[0][k];
        }
        for (k = starts[0]; k < ends[0]; k++) {
            decision += kernels[k] * coefficients[0][k];
        }
        decision += intercepts[0];

        if (decision > 0) {
            return 0;
        }
        return 1;

    }

    double decisions[N_INTERCEPTS];
    double tmp;
    for (i = 0, d = 0, l = N_ROWS; i < l; i++) {
        for (j = i + 1; j < l; j++) {
            tmp = 0.;
            for (k = starts[j]; k < ends[j]; k++) {
                tmp += kernels[k] * coefficients[i][k];
            }
            for (k = starts[i]; k < ends[i]; k++) {
                tmp += kernels[k] * coefficients[j - 1][k];
            }
            decisions[d] = tmp + intercepts[d];
            d = d + 1;
        }
    }

    int votes[N_INTERCEPTS];
    for (i = 0, d = 0, l = N_ROWS; i < l; i++) {
        for (j = i + 1; j < l; j++) {
            votes[d] = decisions[d] > 0 ? i : j;
            d = d + 1;
        }
    }

    int amounts[N_CLASSES];
    for (i = 0, l = N_CLASSES; i < l; i++) {
        amounts[i] = 0;
    }
    for (i = 0; i < N_INTERCEPTS; i++) {
        amounts[votes[i]] += 1;
    }

    int classVal = -1;
    int classIdx = -1;
    for (i = 0; i < N_CLASSES; i++) {
        if (amounts[i] > classVal) {
            classVal = amounts[i];
            classIdx= i;
        }
    }
    return classIdx;

}

int main(int argc, const char * argv[]) {

    /* Features: */
    double features[argc-1];
    int i;
    for (i = 1; i < argc; i++) {
        features[i-1] = atof(argv[i]);
    }

    /* Prediction: */
    printf("%d", predict(features));
    return 0;

}