预处理算法_8_异常值处理

#!/usr/bin/env python
# -*- coding:utf-8 -*-

# <editable>

def execute():
    # <editable>

    import pandas as pd
    import numpy as np
    import warnings
    warnings.filterwarnings('ignore')
    from sqlalchemy import create_engine
    '''
    连接数据库
    '''
    engine = create_engine('mysql+pymysql://root:123123qwe@127.0.0.1:3306/analysis')
    params = {
        "columns": "SUNACTIVITY",
        "method": "Mean_interpolation"

    }
    inputs = {"table": '纯随机性检验'}
    '''
    选择目标数据
    '''
    data_sql = 'select ' + params['columns'] + ' from ' + inputs['table']
    data_in = pd.read_sql_query(data_sql, engine)
    data_in.iloc[1] = np.nan
    print(data_in)
    # data_name = db_utils.query(conn, 'select ' + params['columns'] + ' from ' + inputs['data_in'])
    '''
    找出缺失值
    '''

    def outRange(Ser1):
        QL = Ser1.quantile(float(params['upper_quantile']))
        QU = Ser1.quantile(float(params['lower_quantile']))
        IQR = QU - QL
        Ser1.loc[Ser1 > (QU + 1.5 * IQR)] = None
        Ser1.loc[Ser1 < (QL - 1.5 * IQR)] = None
        return Ser1

    # names = data_name.columns
    # for j in names:
    #     data_in[j] = outRange(data_in[j])

    '''
    对异常值处理
    '''

    '''
    异常值处理方法：
        删除异常值:drop
        中位数插补:Median_interpolation
        众数插补:Mode_interpolation
        均值插补:Mean_interpolation
        线性插值:slinear
        多项式插值:polynomial
    '''
    if params['method'] == 'drop':
        data_out = data_in.dropna()
    elif params['method'] == 'Median_interpolation':
        data_out = data_in.fillna(data_in.median())
    elif params['method'] == 'Mode_interpolation':
        data_out = data_in.fillna(data_in.mode())
    elif params['method'] == 'slinear':
        data_out = data_in.interpolate(method='slinear')
    elif params['method'] == 'quadratic':
        data_out = data_in.interpolate(method='quadratic')
    elif params['method'] == 'Mean_interpolation':
        data_out = data_in.fillna(data_in.mean())
    else:
        data_out = data_in.fillna(data_in.mean())

    '''
    将结果写出
    '''

    print(data_out)

    # </editable>


if __name__ == '__main__':
    execute()