我无法弄清楚 Imshow

Question

所以我在做knn，我的 function 返回一个分类标签（1D），我 plot 它使用散点图得到这个：

我如何将其转换为每行具有“n”个像素的 imshow() plot？

Answer 1

我最近这样做了：

from sklearn.neighbors import KernelDensity

def kde2D(x, y, bandwidth, bins, **kwargs): 
    """Build 2D kernel density estimate (KDE). Adapted from: https://stackoverflow.com/questions/41577705/how-does-2d-kernel-density-estimation-in-python-sklearn-work"""

    # A grid representing the sampled space. Large bins make it faster 
    xx, yy = np.mgrid[df.x.min():df.x.max():bins, 
              df.y.min():df.y.max():bins]

    xy_sample = np.vstack([yy.ravel(), xx.ravel()]).T
    xy_train  = np.vstack([y, x]).T

    # Apply kernel density. Large bandwidth gives more regional effects and is slower
    kde_skl = KernelDensity(bandwidth=bandwidth, **kwargs)
    kde_skl.fit(xy_train)

    # score_samples() returns the kernel density at the grid points linearly
    z = np.exp(kde_skl.score_samples(xy_sample))

    # Reshape it to x,y coordinates corresponding to the original grid
    return np.reshape(z, xx.shape)

# Apply function and plot it
dens = kde2D(df.x, df.y,  bandwidth = 80, bins = 100)

plt.imshow(dens)

干杯，里卡多