使用Logistic回归进行预测始终为0

时间:2019-11-22 23:22:20

标签: machine-learning scikit-learn logistic-regression

我正在尝试根据情感对文本进行分类。不幸的是,我不知道如何正确地使分类器预测任何字符串

global music
def music(layout, list, xs, ys, xp, yp):
    layout.buttons=[]
    for song in list:
        button=Button(text=song, size_hint=(xs, ys), pos_hint={'x':xp, 'y':yp})
        button.bind(on_press=play)
        layout.buttons.append(button)
        layout.add_widget(button)

global play
def play(button):
    song = (button.text)
    try:
        songp.stop()
        print('stop')
    except:
        pass
    finally:
        path = (str(os.getcwd()) + '/Music/' + song + '.mp3')
        songp = SoundLoader.load(str(path))
        songp.play()

每个预测结果都为0,与字符串无关

我得到的全部是。 Yelp数据的准确性:0.7960 亚马逊数据的准确性:0.7960 imdb数据的准确性:0.7487 [0]

我对这个:D很新

0 个答案:

没有答案