visualization for ensemble weights, threshold and f1_score of binary classification

vizual_ensemble.py

import pandas as pd
import numpy as np
from sklearn.metrics import confusion_matrix,f1_score
from bokeh.io import show, output_file
from bokeh.models import (ColumnDataSource,HoverTool,FixedTicker,PrintfTickFormatter)
from bokeh.plotting import figure

def prabtoclass(series,hold):
    #turn probability to binary label by target threshold
    return [1 if i>=hold else 0 for i in series]

def ensemble_weight(data, step, label, input1, input2):
    #data: pandasDataframe; step: number of step to test between 0~1
    #label: the corresct binary label #input1, input2: the two results to ensmble
    weights = np.linspace(0,1,step)
    thresholds = np.linspace(0,1,step)
    x=[]
    y=[]
    z =[]
    for i in weights:
        for j in thresholds:
            x.append(i)
            y.append(j)
            z.append(f1_score(d[label],prabtoclass(d[input1]*i+d[input2]*(1-i),j)))
    return x, y, z

def visual_en(figure_title, x, y, z):
    cource=ColumnDataSource({'x': x, 'y': y, 'z': z})

    toolbar = 'hover,save,pan,box_zoom,reset,wheel_zoom'

    alpha = [r*(1/max(z)) for r in z]

    p = figure(title=figure_title, tools=toolbar, width=800, height=800)
    p.rect('x','y', width=0.02, height=0.02, source=cource,fill_alpha = alpha, line_alpha = alpha)

    hover = p.select_one(HoverTool)
    hover.tooltips = [('weight','@x'),("threshold","@y"),('f1','@f1')]
    p.xaxis.axis_label = 'weight'
    p.yaxis.axis_label = "boolean treshold"

    output_file(figure_title)
    show(p)