Exemplo n.º 1
0
def test_ant():
    # test function on i.antipiev logs
    with open('../logs_ant/result_boosted.txt', 'w') as f:
        for time_window in range(5000, 30001, 5000):
            print('Working with time window equals to {} ms...'.format(time_window), file=f)
            bot_parser = Parser('../logs_ant/log_bot_Game.csv', '../logs_ant/log_bot_Touch.csv', step=time_window)
            human_parser = Parser('../logs_ant/log_human_Game.csv', '../logs_ant/log_human_Touch.csv', step=time_window)
            bot_parser.set_label(1)
            human_parser.set_label(0)
            data = pd.concat([bot_parser.df, human_parser.df], ignore_index=True)
            features, targets = prepare_data(data)
            clf = Classifier(boosted=True).get_data(features, targets, test_size=0.25).fit()
            print('Score:', clf.score(), file=f)
            print('Confusion matrix:\n{}\n'.format(clf.confusion_matrix()), file=f)
Exemplo n.º 2
0
def human_vs_human_3():
    # Human vs human vs human classification
    with open('human_vs_human_3.txt', 'w') as f:
        for time_window in range(5000, 30001, 5000):
            print('Working with {} second window...'.format(time_window))
            h1_logs = Parser('../logs/human_game_logs_1.csv', '../logs/human_touch_logs_1.csv', step=time_window)
            h2_logs = Parser('../logs_new/human/session_2019-01-27_15-01-22/gamelog_Game_2019-01-27_15-01-22.csv',
                             '../logs_new/human/session_2019-01-27_15-01-22/gamelog_Touch_2019-01-27_15-01-22.csv',
                             step=time_window)
            h3_logs = Parser('../logs_new/human/session_2019-01-25_14-10-15/gamelog_Game_2019-01-25_14-10-15.csv',
                             '../logs_new/human/session_2019-01-25_14-10-15/gamelog_Touch_2019-01-25_14-10-15.csv',
                             step=time_window)

            h1_logs.set_label(1)
            h2_logs.set_label(2)
            h3_logs.set_label(3)

            data = pd.concat([h1_logs.df, h2_logs.df, h3_logs.df], ignore_index=True)
            features, targets = prepare_data(data)

            if time_window == 20000:
                scatter(features, targets, save='h_3')
                simple_plot(features, targets, save='h_3')
                tsne_plot(features, targets, perp=10, save='h_3')

            clf = Classifier(boosted=False).get_data(features, targets, test_size=0.25).fit()
            print('Score:', clf.score(), file=f)
            print('Confusion matrix:\n{}\n'.format(clf.confusion_matrix()), file=f)