diff --git a/assignment-1/submission/16307100065/.keep b/assignment-1/submission/16307100065/.keep
new file mode 100644
index 0000000000000000000000000000000000000000..e69de29bb2d1d6434b8b29ae775ad8c2e48c5391
diff --git a/assignment-1/submission/16307100065/README.md b/assignment-1/submission/16307100065/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..06a0098602e802319c84db26985360e4c9e99cfb
--- /dev/null
+++ b/assignment-1/submission/16307100065/README.md
@@ -0,0 +1,214 @@
+# KNN实验探索
+
+## 一.KNN算法代码实现
+
+首先定义KNN类，KNN类中包括三个方法：
+
+```python
+import numpy as np
+from collections import Counter #用于计数
+def __init__(self,k):
+    	#输入超参k，并赋给内部变量self._k;定义另外内部三个重要变量 		         
+        #self._train_data,self._train_label,self._test_data
+        self._k = k
+        self._train_data=None
+        self._train_label = None
+        self._test_data=None
+```
+
+```python
+def fit(self,train_data,train_label):
+    	#输入train_data，train_label并赋给内部变量
+        self._train_data=train_data
+        self._train_label=train_label
+```
+
+```python
+def predict(self,test_data):
+   		#输入test_data并赋给内部变量
+        self._test_data=test_data
+        predicts_ =[]
+        #遍历测试集
+        for i in self._test_data:
+            #对测试集中的数据求与每一个训练集中数据的欧氏距离
+            distances_ = [np.sum((i-x)**2)**0.5 for x in self._train_data]
+            distances = np.array(distances_)
+            #用Counter函数求距离前k个中0-1的个数
+            sorted_distances = np.argsort(distances)
+            topK = [self._train_label[j] for j in sorted_distances[0:self._k]]
+            votes = Counter(topK)
+            #预测结果为距离前k个中0-1数量多的种类
+            predict = votes.most_common(1)[0][0]
+            predicts_.append(predict)
+        predicts = np.array(predicts_)
+        return predicts
+```
+
+
+
+## 二.试验探究
+
+### 1.二维随机正态分布的简单分类实验
+
+（1）两维之间完全不相关
+
+```python
+import numpy as np
+from source import KNN
+import matplotlib.pyplot as plt
+#每一维均值为0，方差为10，并且两维独立，创建1000个数据
+cov = [[10,0],[0,10]]
+data = np.around(np.random.multivariate_normal((0,0),cov,1000),2)
+#对应分类随机取0或1
+label = np.random.choice([0,1],size=1000,replace=True)
+#按8：2的比例分为训练集和测试集
+n = len(data)//5
+train_data = data[0:4*n]
+train_label = label[0:4*n]
+test_data = data[4*n:]
+test_label = label[4*n:]
+#调用KNN类，k赋值5，将训练集输入模型
+model = KNN(5)
+model.fit(train_data, train_label)
+
+#绘制分类图
+#第一维和第二维分别作为x,y轴
+x_show = train_data[:,0]
+y_show = train_data[:,1]
+x_min,x_max=x_show.min(),x_show.max()
+y_min,y_max=y_show.min(),y_show.max()
+#将坐标系分为200×200的网格
+xx,yy = np.meshgrid(np.linspace(x_min,x_max,200),np.linspace(y_min,y_max,200))
+#将网格放入模型预测，预测每一个网格的分类
+z1 = np.c_[xx.ravel(),yy.ravel()]
+z = np.array(z1)
+pred = model.predict(z)
+pred = pred.reshape(xx.shape)
+#绘制网格分类图和训练集的散点图
+plt.pcolormesh(xx,yy,pred,cmap=plt.cm.Pastel1)
+plt.scatter(x_show,y_show,s=80,c=train_label,cmap=plt.cm.spring,edgecolors='k')
+plt.xlim(xx.min(),xx.max())
+plt.ylim(yy.min(),yy.max())
+plt.show()
+```
+
+![输入图片说明](https://images.gitee.com/uploads/images/2021/0329/195821_9a6657ad_8843048.png "二维结果.png")
+
+结果如图所示。由于data是根据二元正态分布随机取值的，并且label也是在0-1之间随机选取的，所以data和label之间是完全无关的。所以分类图也是不规则的。
+
+（2）二维之间完全正相关
+
+```python
+#使两个维度相关系数为1
+cov = [[10,10],[10,10]]
+```
+
+![输入图片说明](https://images.gitee.com/uploads/images/2021/0329/195844_3f64dcf7_8843048.png "二维相关.png")
+
+结果如图所示。由于两维相关系数为1，所以所有点都在y=x的直线上。由于label的随机选取，所以0-1分类区域是在这条直线上的随机分布。
+
+（3）完全负相关
+
+```python
+#两个维度相关系数为-1
+cov = [[10,-10],[-10,10]]
+```
+
+![输入图片说明](https://images.gitee.com/uploads/images/2021/0329/195855_21448894_8843048.png "负二维.png")
+
+结果如图所示。
+
+（4）一般情况下
+
+```python
+#使相关系数为0.2
+cov = [[10,2],[2,10]]
+```
+
+![输入图片说明](https://images.gitee.com/uploads/images/2021/0329/195908_c9a5e2fa_8843048.png "0.2.png")
+
+```python
+#使相关系数为0.8
+cov = [[10,8],[8,10]]
+```
+
+![输入图片说明](https://images.gitee.com/uploads/images/2021/0329/195921_5cf96f45_8843048.png "0.8.png")
+
+分别将相关系数设为0.2与0.8，可以看出0.2时分布还相对散乱，而0.8时已经可以看出明显的线性关系。同样，由于label的随机选取，在直线上0-1的分布仍然是随机的。
+
+### 2.二维随机正态分布的多分类实验
+
+```python
+#相关系数取0.8
+cov = [[10,8],[8,10]]
+#五分类任务
+label = np.random.choice([0,1,2,3,4],size=1000,replace=True)
+```
+
+![输入图片说明](https://images.gitee.com/uploads/images/2021/0329/195937_f438d130_8843048.png "五分类0.8.png")
+
+```python
+#相关系数取0.2
+cov = [[10,2],[2,10]]
+#五分类任务
+label = np.random.choice([0,1,2,3,4],size=1000,replace=True)
+```
+
+![输入图片说明](https://images.gitee.com/uploads/images/2021/0329/195959_868f36e5_8843048.png "五0.2.png")
+
+结果如图所示。与简单分类结果类似，这是由于两维的相关系数是一样的，而仅仅是将label从0-1随机取值变为0-1-2-3-4随机取值，并没有本质变化，只是人为定义的分类数变多了。
+
+### 3. 实验发现
+
+#### （1）简单分类与多分类
+
+以相关系数为0.2的二维正态分布，k值取5为例。
+
+运行10次0-1的简单二分类实验，结果分别为：
+
+acc = 0.51,0.545,0.535,0.515 ,0.435, 0.465,0.56,0.53, 0.515,0.53
+
+平均acc=0.514
+
+运行10次三分类实验，结果如下：
+
+acc = 0.295,0.32,0.35,0.34,0.31,0.405 ,0.325 ,0.3,0.32,0.32
+
+平均acc = 0.329
+
+运行10次四分类实验，结果如下：
+
+acc =  0.25,0.22,0.255,0.19,0.215,0.335,0.3,0.25,0.28,0.225
+
+平均acc = 0.252
+
+运行10此五分类实验，结果如下：
+
+acc = 0.265,0.185,0.22,0.235,0.195,0.19,0.22, 0.23,0.2,0.215
+
+平均acc = 0.2155
+
+
+
+可以发现，当分类数越多时，预测的准确率越低。这应该是由于data和label本身之间没有相关性，acc近似等于（1/分类数），和瞎猜的准确率是近似的。
+
+#### （2）k值的选取
+
+以简单二分类为例。
+
+k值取5，运行10次的平均acc =0.512
+
+k值取4，运行10次的平均acc =0.507
+
+k值取6，运行10次的平均acc =0.485
+
+k值取3，运行10次的平均acc =0.495
+
+k值取7，运行10次的平均acc =0.508
+
+
+
+可知，k取5是恰当的。但同样由于data和label之间没有相关性，所以不同k值之间准确率的差异不大，都在0.5左右，近似于瞎猜。因为分类区域本身就是随机的，距离预测点越近的点并不代表属于该点类别的概率越大。
+
+所以，该次尝试的调参是没有意义的。如果选取本身有意义，并且属性与分类有关系的数据组进行测试，取不同的k值的准确率才会有显著差异，此时的调参才有意义。
+
diff --git a/assignment-1/submission/16307100065/img/.keep b/assignment-1/submission/16307100065/img/.keep
new file mode 100644
index 0000000000000000000000000000000000000000..e69de29bb2d1d6434b8b29ae775ad8c2e48c5391
diff --git a/assignment-1/submission/16307100065/img/0.2.png b/assignment-1/submission/16307100065/img/0.2.png
new file mode 100644
index 0000000000000000000000000000000000000000..5cc12d92a5e696a6181dc47fef6caf9de5a4d7c9
Binary files /dev/null and b/assignment-1/submission/16307100065/img/0.2.png differ
diff --git a/assignment-1/submission/16307100065/img/0.8.png b/assignment-1/submission/16307100065/img/0.8.png
new file mode 100644
index 0000000000000000000000000000000000000000..abe0d2e4626cc82b2e44961f668c4d732aa057cf
Binary files /dev/null and b/assignment-1/submission/16307100065/img/0.8.png differ
diff --git "a/assignment-1/submission/16307100065/img/\344\272\214\347\273\264\347\233\270\345\205\263.png" "b/assignment-1/submission/16307100065/img/\344\272\214\347\273\264\347\233\270\345\205\263.png"
new file mode 100644
index 0000000000000000000000000000000000000000..4eb7d1805b16224826b6522b18680a6951f08085
Binary files /dev/null and "b/assignment-1/submission/16307100065/img/\344\272\214\347\273\264\347\233\270\345\205\263.png" differ
diff --git "a/assignment-1/submission/16307100065/img/\344\272\214\347\273\264\347\273\223\346\236\234.png" "b/assignment-1/submission/16307100065/img/\344\272\214\347\273\264\347\273\223\346\236\234.png"
new file mode 100644
index 0000000000000000000000000000000000000000..cfb88bc7cf41bc6bb02cd3467b2f475c43a405a2
Binary files /dev/null and "b/assignment-1/submission/16307100065/img/\344\272\214\347\273\264\347\273\223\346\236\234.png" differ
diff --git "a/assignment-1/submission/16307100065/img/\344\272\2240.2.png" "b/assignment-1/submission/16307100065/img/\344\272\2240.2.png"
new file mode 100644
index 0000000000000000000000000000000000000000..5d151517b7b0da7e0ccf89ae43a44b848831e19c
Binary files /dev/null and "b/assignment-1/submission/16307100065/img/\344\272\2240.2.png" differ
diff --git "a/assignment-1/submission/16307100065/img/\344\272\224\345\210\206\347\261\2730.8.png" "b/assignment-1/submission/16307100065/img/\344\272\224\345\210\206\347\261\2730.8.png"
new file mode 100644
index 0000000000000000000000000000000000000000..d5d72a9445ffdab5b80e24d02ee76a3789244d50
Binary files /dev/null and "b/assignment-1/submission/16307100065/img/\344\272\224\345\210\206\347\261\2730.8.png" differ
diff --git "a/assignment-1/submission/16307100065/img/\350\264\237\344\272\214\347\273\264.png" "b/assignment-1/submission/16307100065/img/\350\264\237\344\272\214\347\273\264.png"
new file mode 100644
index 0000000000000000000000000000000000000000..5f28d4a81b3deaa217a3d1a78602d6df915b3226
Binary files /dev/null and "b/assignment-1/submission/16307100065/img/\350\264\237\344\272\214\347\273\264.png" differ
diff --git a/assignment-1/submission/16307100065/source.py b/assignment-1/submission/16307100065/source.py
new file mode 100644
index 0000000000000000000000000000000000000000..4ae97853ca5174d7f23998d209a5603a4a7b7c58
--- /dev/null
+++ b/assignment-1/submission/16307100065/source.py
@@ -0,0 +1,35 @@
+#算法代码
+import numpy as np
+from collections import Counter
+class KNN:
+
+    def __init__(self,k):
+        self._k = k
+        self._train_data=None
+        self._train_label = None
+        self._test_data=None
+        
+    def fit(self,train_data,train_label):
+        self._train_data=train_data
+        self._train_label=train_label
+        
+        
+    def predict(self,test_data):
+        self._test_data=test_data
+        predicts_ =[]
+        #遍历测试集
+        for i in self._test_data:
+            #对测试集中的数据求距离每一个训练集中数据的欧氏距离
+            distances_ = [np.sum((i-x)**2)**0.5 for x in self._train_data]
+            distances = np.array(distances_)
+            #用Counter函数求距离前k个
+            sorted_distances = np.argsort(distances)
+            topK = [self._train_label[j] for j in sorted_distances[0:self._k]]
+            votes = Counter(topK)
+            #预测结果
+            predict = votes.most_common(1)[0][0]
+            predicts_.append(predict)
+        predicts = np.array(predicts_)
+        return predicts
+    
+
diff --git a/assignment-1/submission/16307100065/train&test.py b/assignment-1/submission/16307100065/train&test.py
new file mode 100644
index 0000000000000000000000000000000000000000..31f28e4c3580e54f001d4ac69fb1c3f9ddf533a9
--- /dev/null
+++ b/assignment-1/submission/16307100065/train&test.py
@@ -0,0 +1,39 @@
+#实验代码
+import numpy as np
+from lknn import KNN
+import matplotlib.pyplot as plt
+#每一维均值为0，方差为10，并且两维独立，创建1000个数据
+cov = [[10,0],[0,10]]
+data = np.around(np.random.multivariate_normal((0,0),cov,1000),2)
+#对应分类随机取0或1
+label = np.random.choice([0,1],size=1000,replace=True)
+#按8：2的比例分为训练集和测试集
+n = len(data)//5
+train_data = data[0:4*n]
+train_label = label[0:4*n]
+test_data = data[4*n:]
+test_label = label[4*n:]
+#调用KNN类，k赋值5，将训练集输入模型
+model = KNN(5)
+model.fit(train_data, train_label)
+#绘制分类图
+#第一维和第二维分别作为x,y轴
+x_show = train_data[:,0]
+y_show = train_data[:,1]
+x_min,x_max=x_show.min(),x_show.max()
+y_min,y_max=y_show.min(),y_show.max()
+xx,yy = np.meshgrid(np.linspace(x_min,x_max,200),np.linspace(y_min,y_max,200))
+#将网格放入模型预测，预测每一个网格的分类
+z1 = np.c_[xx.ravel(),yy.ravel()]
+z = np.array(z1)
+pred = model.predict(z)
+pred = pred.reshape(xx.shape)
+#绘制网格分类图和训练集的散点图
+plt.pcolormesh(xx,yy,pred,cmap=plt.cm.Pastel1)
+plt.scatter(x_show,y_show,s=80,c=train_label,cmap=plt.cm.spring,edgecolors='k')
+plt.xlim(xx.min(),xx.max())
+plt.ylim(yy.min(),yy.max())
+plt.show()
+#计算acc
+res = model.predict(test_data)
+print("acc =",np.mean(np.equal(res, test_label)))
\ No newline at end of file