Question

我最近查看了Creating interactive applications and toolbars in Jupyter Notebook上的mtap主题，并决定播放一个示例，其中作者编写了一个用于研究推文的界面，但不幸的是，输出出现矩阵混淆的部分遇到了问题< / p>

Figure (axes = [Axis (label = 'Actual', label_offset = '4em', orientation = 'vertical', scale = OrdinalScale (reverse = True) ...

不幸的是，该视频并不太了解我所错过的内容，请您帮忙实现此界面。该视频大约在17分钟左右开始。

我的一段代码。

    confusion_mat = confusion_matrix(y_test, y_pred_bigram)


    label_map = {0.0 : 'Negative', 1.0 : 'Positive'}
    color_map = {0.0 : 'red', 1.0 : 'limegreen'}

    row_scale = OrdinalScale(reverse=True)
    col_scale = OrdinalScale()
    color_scale = ColorScale(scheme='Greens')
    row_axis = Axis(scale=row_scale, orientation='vertical', label='Actual', 
    label_offset='4em')
    col_axis = Axis(scale=col_scale, label='Predicted')
    conf_mat_grid = GridHeatMap(
            column=['Negative', 'Positive'],
            row=['Negative', 'Positive'],
            color=np.log(confusion_mat),
            scales={'row' : row_scale, 'column' : col_scale, 'color' : 
    color_scale},
            interactions={'click' : 'select'},
            anchor_style={'stroke' : 'red', 'strocke-width' : 3},
            selected_style={'stroke' : 'red'})

    y, x, text = zip(*[(label_map[i], label_map[j],
                      '{:,}'.format(confusion_mat[i, j])) for i in range(2) 
    for j in range(2)])

    grid_labels = Label(x=x, y=y, text=text,
                       scales={'x' : col_scale, 'y' : row_scale},
                       font_size=26,
                       font_weight='bolder')

     grid_fig = Figure(title='Confusion Matrix',
                      marks=[conf_mat_grid, grid_labels],
                      padding_y=0.0,
                      axes=[row_axis, col_axis],
                      fig_margin=dict(left=80, top=80, bottom=40, right=100))
    grid_fig.layout.width = '460px'
    grid_fig.layout.height = '400px'
    grid_fig

在Jupyter Notebook中构建交互式应用程序和仪表板

0 个答案: