2022-11-22-17:07:25

Author: lifeiyang

leetcode/editor/cn/[28]找出字符串中第一个匹配项的下标.py

@@ -1,4 +1,58 @@
+import string
+
+
 # leetcode submit region begin(Prohibit modification and deletion)
 class Solution:
+    def build_kmp(self, needle):
+        # 定义 dp[i][j] 为: 在第 i 个位置，遇到 j 字母，应该去的位置
+        dp = {}
+        for i in range(len(needle)):
+            dp[i] = {}
+            for j in string.ascii_lowercase:
+                dp[i][j] = 0
+        
+        # 基础情况，在第 0 个位置，遇到第 0 个字母，应该跳转到第 1 个位置
+        dp[0][needle[0]] = 1
+        
+        # X 为跟随在 i 后边的重叠子问题状态
+        X = 0
+        for i in range(1, len(needle)):
+            for j in string.ascii_lowercase:
+                if j == needle[i]:
+                    dp[i][j] = i + 1
+                else:
+                    dp[i][j] = dp[X][j]
+            X = dp[X][needle[i]]
+        return dp
+    
+    def strStrKMP(self, haystack: str, needle: str) -> int:
+        kmp_pattern_dp = self.build_kmp(needle)
+        m, n = len(needle), len(haystack)
+        j = 0
+        for idx, char in enumerate(haystack):
+            j = kmp_pattern_dp[j][char]
+            if j == m:
+                return idx - m + 1
+        
+        return -1
+    
     def strStr(self, haystack: str, needle: str) -> int:
+        res = -1
+        m, n = len(haystack), len(needle)
+        if m < n:
+            return res
+        
+        for idx in range(m - n + 1):
+            if haystack[idx:idx + n] == needle:
+                return idx
+        return res
+
+
 # leetcode submit region end(Prohibit modification and deletion)
+
+
+if __name__ == "__main__":
+    solution = Solution()
+    print(solution.strStr("a", "a"), 0)
+    print(solution.strStr("sadbutsad", "sad"), 0)
+    print(solution.strStr("leetcode", "leeto"), -1)

machine_learning_algorithm/__init__.py

@@ -0,0 +1,9 @@
+# !/usr/bin/python
+# -*- coding: utf-8 -*-
+#
+# Copyright @2022 AI, ZHIHU Inc. (zhihu.com)
+#
+# @author: lifeiyang <[email protected]>
+# @date: 2022/11/24
+#
+""""""

machine_learning_algorithm/kmeans/README.md

@@ -0,0 +1,13 @@
+# 实现 kmeans
+
+输入数据为：
+
+data_list = [[0,5],[0,6],[4,0],[5,0]]
+k =2
+
+对应图为：
+
+![](https://cdn.lifeiyang.cn/img_2022/202211241634692.png)
+
+输出的结果为:
+[[0,5],[0,6]] 和 [[4,0],[5,0]]

machine_learning_algorithm/kmeans/__init__.py

@@ -0,0 +1,9 @@
+# !/usr/bin/python
+# -*- coding: utf-8 -*-
+#
+# Copyright @2022 AI, ZHIHU Inc. (zhihu.com)
+#
+# @author: lifeiyang <[email protected]>
+# @date: 2022/11/24
+#
+""""""

machine_learning_algorithm/kmeans/kmeans.py

@@ -0,0 +1,69 @@
+# !/usr/bin/python
+# -*- coding: utf-8 -*-
+#
+# Copyright @2022 AI, ZHIHU Inc. (zhihu.com)
+#
+# @author: lifeiyang <[email protected]>
+# @date: 2022/11/24
+#
+""""""
+import math
+
+
+class Solution:
+    def distance(self, node1, node2):
+        return math.sqrt(math.pow(node1[0] - node2[0], 2) + math.pow(node1[1] - node2[1], 2))
+    
+    def point_mean(self, point_list):
+        point_x = sum([point[0] for point in point_list]) / len(point_list)
+        point_y = sum([point[1] for point in point_list]) / len(point_list)
+        return (point_x, point_y)
+    
+    def kmeans(self, data_list, k):
+        # 1、随机初始化点
+        k_cluster = {}
+        for i in range(k):
+            k_cluster[tuple(data_list[i])] = []
+        
+        # 2、根据距离加入质心
+        for point in data_list:
+            min_distance = math.inf
+            min_kernel = None
+            for kernel in k_cluster:
+                if min_distance > self.distance(point, kernel):
+                    min_distance = self.distance(point, kernel)
+                    min_kernel = kernel
+            
+            k_cluster[min_kernel].append(point)
+        
+        # 3、开始循环迭代
+        k_kernel_old = k_cluster.copy()
+        while True:
+            # 新一轮的迭代
+            # 1、寻找当前的质心
+            k_cluster = {}
+            for kernel in k_kernel_old:
+                kernel_mean = self.point_mean(k_kernel_old[kernel])
+                k_cluster[kernel_mean] = []
+            
+            # 2、根据距离加入质心
+            for point in data_list:
+                min_distance = math.inf
+                min_kernel = None
+                for kernel in k_cluster:
+                    if min_distance > self.distance(point, kernel):
+                        min_distance = self.distance(point, kernel)
+                        min_kernel = kernel
+                k_cluster[min_kernel].append(point)
+            
+            if k_cluster == k_kernel_old:
+                print("final kmeans:", k_cluster)
+                break
+            else:
+                k_kernel_old = k_cluster.copy()
+                print("now kmeans:", k_cluster)
+
+
+if __name__ == "__main__":
+    solution = Solution()
+    print(solution.kmeans([[0, 5], [0, 6], [4, 0], [5, 0]], 2))