C++实现二维数组按行乱序（UCI iris数据集整理）

在写BP神经网络时候碰到的小问题：UCI iris数据集是按顺序排的，我取的是前130行训练，后20行测试。因为到最后训练也是第三类，测试也都是第三类，是对于第三类的局部最优解，所以最后不用几轮训练就能达到>0.9的准确率，但全局上就不对。

于是写了个小程序随机改数据集的训练顺序。
因为只改行的顺序，对应的列不变。思想是用random的shuffle函数随机改行号，再把后面的列对应过去，以此实现乱序。
UCI iris数据集：
https://download.csdn.net/download/pvfeldt/18123868?spm=1001.2014.3001.5503

代码
下面是代码，有些地方写的不是很好，比如实现读文件时候数据和分类分开存读了两遍，如果是开个全局变量或者搞个class就可以避免；或者是这个想法本身就有点麻烦，水平有限见谅。

#include 
#include 
#include 
#include 
#include 
#include 
using namespace std;

//read file,store 4 feature columns in 2D double array;
double** irisData(string filename, int lines)
{
     
    ifstream file(filename, ios::binary);
    double** irisData = new double* [lines];
    for (int i = 0; i < lines; i++)
    {
     
        irisData[i] = new double[4];
    }
    for (int i = 0; i < lines; i++)
    {
     
        double numTmp = 0;
        char commaTmp = ' ';
        string classificationTmp = " ";
        for (int j = 0; j < 4; j++)
        {
     
            file >> numTmp >> commaTmp;
            irisData[i][j] = numTmp;
        }
        file >> classificationTmp;
    }
    file.close();
    return irisData;
}

//read file,store classification column in 1D string array;
string* irisClassification(string filename, int lines)
{
     
    ifstream file(filename, ios::binary);
    string* irisClassification = new string[lines];
    for (int i = 0; i < lines; i++)
    {
     
        double numTmp = 0;
        char commaTmp = ' ';
        string classificationTmp = " ";
        for (int j = 0; j < 4; j++)
        {
     
            file >> numTmp >> commaTmp;
        }
        file >> classificationTmp;
        irisClassification[i] = classificationTmp;
    }
    file.close();
    return irisClassification;
}

//randomly change the line index;
vector<int> changeLineIndexOrder(int lines)
{
     
    vector<int> lineIndex(lines);
    for (int i = 0; i < lines; i++)
    {
     
        lineIndex[i] = i;
    }
    unsigned seed = std::chrono::system_clock::now().time_since_epoch().count();
    shuffle(lineIndex.begin(), lineIndex.end(), default_random_engine(seed));
    return lineIndex;
}

//change feature array according to the change of line index;
double** changeDataOrder(double** irisD, vector<int> lineI, int lines)
{
     
    double** newIrisData = new double* [lines];
    for (int i = 0; i < lines; i++)
    {
     
        newIrisData[i] = new double[4];
    }
    for (int i = 0; i < lines; i++)
    {
     
        for (int j = 0; j < 4; j++)
        {
     
            newIrisData[i][j] = irisD[lineI[i]][j];
        }
    }
    return newIrisData;
}

//change classification array according to the change of line index;
string* changeClassificationOrder(string* irisC, vector<int> lineI, int lines)
{
     
    string* newIrisClassification = new string[lines];
    for (int i = 0; i < lines; i++)
    {
     
        newIrisClassification[i] = irisC[lineI[i]];
    }
    return newIrisClassification;
}

int main()
{
     
    int lines = 150;
    double** irisDataReading = irisData("iris.data", lines);
    string* irisClassificationReading = irisClassification("iris.data", lines);
    vector<int> lineIndex = changeLineIndexOrder(lines);
    double** newIrisDataReading = changeDataOrder(irisDataReading, lineIndex, lines);
    string* newIrisClassificationReading = changeClassificationOrder(irisClassificationReading, lineIndex, lines);

    cout << "original data sequence:" << endl;
    for (int i = 0; i < lines; i++)
    {
     
        cout << i << "  ";
        for (int j = 0; j < 4; j++)
        {
     
            cout << irisDataReading[i][j] << "  ";
        }
        cout << irisClassificationReading[i] << endl;
    }
    
    cout << "changed data sequence:" << endl;
    for (int i = 0; i < lines; i++)
    {
     
        cout << lineIndex[i] << "  ";
        for (int j = 0; j < 4; j++)
        {
     
            cout << newIrisDataReading[i][j] << "  ";
        }
        cout << newIrisClassificationReading[i] << endl;
    }


    for (int i = 0; i < lines; i++)
    {
     
        delete[] irisDataReading[i];
        delete[] newIrisDataReading[i];
    }
    delete[] irisDataReading;
    delete[] newIrisDataReading;
    delete[] irisClassificationReading;
    delete[] newIrisClassificationReading;

    system("pause");
    return 0;
}

Over.