外星人入侵

最简单易懂的聚类方法，用python实现k-means对数据的聚类(The most straightforward clustering method, using python to achieve k-means clustering of data)

PSO 算法属于进化算法的一种，和模拟退火算法相似，它也是从随机解出发，通过迭代寻找最优解，它也是通过适应度来评价解的品质，但它比遗传算法规则更为简单，它没有遗传算法的“交叉”(Crossover) 和“变异”(Mutation) 操作，它通过追随当前搜索到的最优值来寻找全局最优。给出一个用pso来解决的寻优问题

用于钢筋混凝土结构构件在Opensees中的模拟(INF simulation of RC columns in Opensees)

说明：  斯坦福大学自动驾驶城市挑战赛的代码，比较值得一看(Code of Stanford University auto driving city challenge)

采用了scrapy框架爬取链家网二手房的数据，可以导入到数据库中保存数据(Using the scrapy frame to climb the data of the second-hand house of the chain home network, it can be imported into the database to save the data)

说明：  包含unet/google-v2/CNN等多种神经网络的模型(Multiple Neural Network Models)

smote过采样方法对不平衡数据集进行人工数据合成(smote oversampling and classification)

说明：  是最基础的图像压缩，是深度学习的入门教程，适合初学者进行学习(It is the most basic image compression, an introductory course for deep learning, suitable for beginners to learn.)

说明：  DBSCAN（Density-Based Spatial Clustering of Applications with Noise） 为一种基于密度的聚类算法，它不仅可以找出具有任何形状的簇，而且还可以用于检测离群值。其基本思想为数据点分布紧凑的应被划分为一类，而周围未分布有或仅有极少数点的数据点则有可能为离群值。通过python实现了该聚类方法的计算。(DBSCAN (density based spatial clustering of applications with noise) is a density based clustering algorithm. It can not only find clusters with any shape, but also detect outliers. The basic idea is that the data points with compact distribution should be divided into a class, while the data points with few or no distribution around may be outliers. This clustering method is implemented by python.)

说明：  将物体用一堆点框出来，实际先初始化一堆点，然后迭代靠近物体边缘最终框出物体。(The object is framed with a pile of points. In fact, first initialize a pile of points, then iterate close to the edge of the object, and finally frame the object.)