diff --git a/Exercise1/Exercise1.py b/Exercise1/Exercise1.py
new file mode 100644
index 0000000..3602479
--- /dev/null
+++ b/Exercise1/Exercise1.py
@@ -0,0 +1,66 @@
+import matplotlib.pyplot as plt
+import numpy as np
+import xlrd
+import xlsxwriter
+import math
+import random
+
+ExcelFile=xlrd.open_workbook(r'C:\Users\Aurora\Desktop\data.xlsx')
+data=ExcelFile.sheet_by_index(0)
+
+
+size=569
+X = [[0 for i in range(30)] for j in range(size)]
+W = [random.uniform(-0.3,0.3) for i in range(31)]  ##随机生成权值W
+Y = [0 for i in range(size)]
+dL = [0 for i in range(size)]  ##dL/dz
+alpha=0.001
+
+
+for i in range (1,round(0.7*size)):  ###初始化输入X
+    for j in range(2,32):
+       X[i-1][j-2]=data.cell_value(i,j)
+
+def sigmoid(x):
+    return 1/(1+math.exp(-x))
+
+n=80  ##训练次数n
+while(n):
+  sumdL=0
+
+  for i in range (0,round(0.7*size)-1):  ##用70%训练模型
+      z=0                           ##z暂时储存每一次的wx，最后相加得到总的W*X
+      y = data.cell_value(i + 1, 1)
+
+      for j in range(0,30):
+          z=z+W[j]*X[i][j]+W[30]  ##z=wx+b, b为W[30]
+
+      Y[i]=sigmoid(z)
+      try:
+          dL[i]=-y/(1+math.exp(z)) + (1-y)/(1+math.exp(-z))
+      except OverflowError:
+          dL[i]=-y/(1+math.exp(z/10)) + (1-y)/(1+math.exp(-z/10)) ##防止计算exp发生溢出
+
+      sumdL=sumdL+dL[i]   ##dL求和
+
+  for i in range(0,30):
+      W[i] = W[i] - alpha/30*sumdL
+
+  n=n-1
+
+correctNum=0
+for k in range (round(0.7*size),size-1):  ##用另外30%测试模型
+    z=0                           ##z暂时储存每一次的wx，最后相加得到总的W*X
+    y = data.cell_value(k + 1, 1)
+    for j in range(0,30):
+        z=z+W[j]*X[k][j]+W[30]  ##z=wx+b, b为W[30]
+    Y[k]=sigmoid(z)
+
+    if ((y-Y[k])*(y-Y[k])<0.25):  ##如果输出Y在0.5以下，认为是0类，在0.5以上，为1类
+
+        correctNum+=1
+
+print('测试样本正确率: ',correctNum/(0.3*size))
+
+
+
diff --git a/Exercise1/Result.png b/Exercise1/Result.png
new file mode 100644
index 0000000..176fe7f
Binary files /dev/null and b/Exercise1/Result.png differ
diff --git a/Exercise1/data.xlsx b/Exercise1/data.xlsx
new file mode 100644
index 0000000..1c343b3
Binary files /dev/null and b/Exercise1/data.xlsx differ
diff --git a/Exercise1/readme.md b/Exercise1/readme.md
new file mode 100644
index 0000000..844bdcd
--- /dev/null
+++ b/Exercise1/readme.md
@@ -0,0 +1,7 @@
+Exercise1 is about logistic regression on ``Breast Cancer Wisconsin (Diagnostic) Data Set''.
+
+Here are the data and program files.
+
+Note: 1.我修改了data.xlsx，M类替换成了1，B类替换成了0，这样方便做梯度下降。
+
+2.对70%数据做训练，30%数据做测试，测试正确率为81%左右
diff --git a/README.md b/README.md
index 07ea68d..627c3d7 100644
--- a/README.md
+++ b/README.md
@@ -1,3 +1,4 @@
 # SES2020spring
 
 ## Another day
+NOTE:项目实践的代码放在了SES2020spring了，不是这个SES2020spring-1. 机器学习的Exercise放在这里，因为可以这个SES2020spring-1是folked，可以作pull request