Mooplab(新网银行杯)

Mooplab(新网银行杯)

一、pycharm第三方库安装不上的问题（2020.10.19-10.21）

相关方法：

1. 在终端上安装包
   1. win+R
   2. 输入cmd进入终端
   3. pip install （numpy）‘要装的安装包’ -i https://pypi.tuna.tsinghua.edu.cn/simple some-package
2. 更新pip：终端输入python -m pip install --upgrade pip
3. PyCharm安装教程（Windows):https://www.runoob.com/w3cnote/pycharm-windows-install.html

二、‘mbcs’ codec can’t encode characters in position 0–1: invalid character 问题（2020.10.22）

方法：

1. 原因：路径问题,将字符串作为路径直接使用，需要转义

   出错代码：pd.read_csv(‘duringopereation\patients\patients_csv\23483949.csv’, encoding = “gb2312”)

   修改代码：pd.read_csv(‘duringopereation\patients\patients_csv\23483949.csv’, encoding = “gb2312”)

2. 将路径删掉重新手敲一遍解决（之前是文件属性里复制的路径）

解决方案：

• 用第二个方案解决

三、KNN&SVM(分类方法） （2020.10.22）

import pandas as pd
import numpy as np
from sklearn.neighbors import KNeighborsClassifier

# 读取数据
dfx_train = pd.read_csv(r"E:/新网银行杯/data_b_train.csv")
dfx_test = pd.read_csv(r"E:/新网银行杯/data_b_test.csv")
dfy_train = pd.read_csv(r"E:/新网银行杯/y_train.csv")
print(dfx_train)

# 字典中的key转换为列表
key_value = list(dfx_train.keys())
print('字典中的key转换为列表：', key_value)
x1_train = np.array(list(dfx_train['x_num_0']))
# x2_train = np.array(list(dfx_train['x_num_1']))
x3_train = np.array(list(dfx_train['x_num_2']))
# x4_train = np.array(list(dfx_train['x_num_3']))
# x5_train = np.array(list(dfx_train['x_num_4']))
# x6_train = np.array(list(dfx_train['x_num_5']))
# x7_train = np.array(list(dfx_train['x_num_6']))
# x8_train = np.array(list(dfx_train['x_num_7']))

x_train=np.array([x1_train,x3_train])
x_train=x_train.T
y_train  = np.array(list(dfy_train['target']))
y_train=y_train.T

#
x1_test = np.array(list(dfx_test['x_num_0']))
# x2_test = np.array(list(dfx_test['x_num_1']))
x3_test = np.array(list(dfx_test['x_num_2']))
# x4_test = np.array(list(dfx_test['x_num_3']))
# x5_test = np.array(list(dfx_test['x_num_4']))
x_test=np.array([x1_test,x3_test])
x_test=x_test.T

#==================================
# KNN
#==================================
knn = KNeighborsClassifier()
knn.fit(x_train,y_train)
y_test=knn.predict(x_test)
print(len(y_test))
#==================================
# SVM
#==================================
# from sklearn.svm import SVC
# model = SVC(kernel='rbf', probability=True)#probability=False时，没办法调用 model.predict_proba()函数
# model.fit(x_train,y_train)
# pre = model.predict_proba(C)
# print(pre)
# y_test = model.predict(x_test)
# print(y_test)

#==================================
# 导出
#==================================
data_df = pd.DataFrame(y_test)

# change the index and column name
data_df.columns = ['target']
data_df.index = list(range(0,5767))

# create and writer pd.DataFrame to excel
writer = pd.ExcelWriter('E:/新网银行杯/y_test.xlsx')
data_df.to_excel(writer,'page_1',float_format='%.5f') # float_format 控制精度
writer.save()

四、随机生成数（2020.10.22）

import numpy as np
import pandas as pd
import random
a=[]
for i in range(1,5767):
    a.append(random.randint(0,1))

a = np.array(a)
# prepare for data
data = a.reshape((5766,1))
print(data)
data_df = pd.DataFrame(data)

# change the index and column name
data_df.columns = ['target']
data_df.index = list(range(1,5767))

# create and writer pd.DataFrame to excel
writer = pd.ExcelWriter('Save_Excel.xlsx')
data_df.to_excel(writer,'page_1',float_format='%.5f') # float_format 控制精度
writer.save()

五、基础理论（2020.10.23）

建模

1. read读数据 (读入scv.文件)
2. operation 操作(pre-processing [用pandas])
3. modeling建模
4. result得出结果
5. write写论文

六、第二次开会（2020.10.23）

任务：

1. 数据预处理（-99用取均值的方法）
2. 感知机分类
3. 筛选数据（可手动删除部分信息）
4. 行为信息数据（给时间点加权）

#pnn感知器分类
from sklearn.linear_model import Perceptron
ppn = Perceptron()
ppn.fit(x_train, y_train)
y_test = ppn.predict(x_test)

学习目标：

python:

1. 语法
2. numpy
3. pandas