YiFraternity / OS-K-Means
现有聚类算法面向高维稀疏数据多未考虑类簇可重叠和离群点的存在,导致聚类效果不理想。针对此,提出一种可重叠子空间K-Means聚类算法(An Overlapping Subspace K-Means Clustering Algorithm, OS-K-Means)。给出类簇子空间计算策略,在聚类过程中动态更新每个类簇的属性子空间,并定义合理的约束函数指导聚类过程,从而实现类簇的可重叠性与寻找离群点的效果。具体地,定义合理的目标函数对传统的K-Means算法进行修正,利用熵权约束分别计算每个类簇中每个维度的权重,使用权重值来标识对不同类簇中维度的相对重要性,并加入对重叠程度和离群值数量控制的参数。
☆30Updated 5 years ago
Alternatives and similar repositories for OS-K-Means
Users that are interested in OS-K-Means are comparing it to the libraries listed below
Sorting:
- 数据预处理之缺失值处理,特征选择☆21Updated 6 years ago
- Affinity Propagation Clustering with DTW distance on temporal sequence classification☆20Updated 6 years ago
- 由时间空间成对组成的轨迹序列,通过循环神经网络lstm,自编码器auto-encode,时空密度聚类st-dbscan做异常检测☆71Updated 5 years ago
- This is an implementation of the paper on "Improved K-means algorithm based on density Canopy".☆30Updated 5 years ago
- 常用的特征选择方法☆68Updated 2 years ago
- 改进的k-prototypes聚类算法☆18Updated 4 years ago
- Optimizing k-means++ initialization using PSO☆17Updated 9 years ago
- Python Code For 'Clustering By Fast Search And Find Of Density Peaks' In Science 2014.(原算法地址:https://github.com/lanbing510/DensityPeakClu…☆16Updated 6 years ago
- 机器学习的特征工程,包括特征抽取、特征预处理、特征选择、特征降维。☆25Updated 6 years ago
- Oversampling method based on relative density☆12Updated 4 years ago
- Removal of information from co-association matrix for clustering ensemble☆17Updated 6 years ago
- 时间序列ARIMA模型的销量预测☆63Updated 6 years ago
- 本项目开发了一个机器学习和深度学习的训练工具。该训练工具基于sklearn和pytorch,不仅支持常规训练、交叉验证训练,还支持贝叶斯搜索参数,并可随时自动保存训练模型和日志。☆12Updated last year
- The code of the AAAI-19 paper "AFS: An Attention-based mechanism for Supervised Feature Selection".☆45Updated 6 years ago
- AutoEncoder implements by keras. Including AE, DAE, DAE_CNN, VAE, VAE_CNN, CVAE, Sparse AE, Stacked DAE.☆41Updated 5 years ago
- Source code of our paper: An overlapping community detection algorithm based on density peaks☆13Updated 7 years ago
- 时间序列异常检测☆53Updated 5 years ago
- Code for *Unsupervised Anomaly Detection for Intricate KPIs via Adversarial Training of VAE*☆24Updated 5 years ago
- 聚类算法k-means的简单实现☆39Updated 6 years ago
- 基于KNN聚类算法结合Dynamic Time Warping(动态时间调整)的时间序列分类☆61Updated 5 years ago
- 利用时间序列预测汽车销量☆40Updated 6 years ago
- 数据特征工程、各种机器学习回归模型、回归数据预处理☆44Updated 5 years ago
- 基于Keras框架,结合LSTM/GRU/Arima/WNN实现多方式的水质参数预测☆22Updated 6 years ago
- 使用 tensorflow2.0 实现图卷积神经网络GCN☆20Updated 4 years ago
- 建立SARIMA-LSTM混合模型预测时间序列问题。以PM2.5值为例,使用UCI公开的自2013年1月17日至2015年12月31日五大城市PM2.5小时检测数据,将数据按时间段划分,使用SARIMA过滤其线性趋势,再对过滤后的残差使用LSTM进行预测,最后对预测结果进行…☆79Updated 6 years ago
- 基于深度学习的溶解氧时间序列预测模型☆28Updated 5 years ago
- 基于Python实现了K-Means、GMM、DBSCAN、AGNES等四种常见的聚类算法☆69Updated 6 years ago
- 基于遗传算法的特征选择☆128Updated 5 years ago
- TensorFlow Probability;Time series model☆126Updated 3 years ago
- Transfer Learning JDA and TrAdaboost☆64Updated 7 years ago