Learning certifiably optimal rule lists for categorical data
Learning certifiably optimal rule lists for categorical data
Elaine Angelino (UC Berkeley)
Nicholas Larus-Stone (Harvard)
Daniel Alabi (Harvard)
Margo Seltzer (Harvard University)
Cynthia Rudin (Duke)
We present the design and implementation of a custom discrete optimization technique for building rule lists over a categorical feature space. Our algorithm provides the optimal solution, with a certificate of optimality. By leveraging algorithmic bounds, efficient data structures, and computational reuse, we achieve several orders of magnitude speedup in time and a massive reduction of memory consumption. We demonstrate that our approach produces optimal rule lists on practical problems in seconds. This framework is a novel alternative to CART and other decision tree methods.
More on http://www.kdd.org/kdd2017/
Видео Learning certifiably optimal rule lists for categorical data канала KDD2017 video
Elaine Angelino (UC Berkeley)
Nicholas Larus-Stone (Harvard)
Daniel Alabi (Harvard)
Margo Seltzer (Harvard University)
Cynthia Rudin (Duke)
We present the design and implementation of a custom discrete optimization technique for building rule lists over a categorical feature space. Our algorithm provides the optimal solution, with a certificate of optimality. By leveraging algorithmic bounds, efficient data structures, and computational reuse, we achieve several orders of magnitude speedup in time and a massive reduction of memory consumption. We demonstrate that our approach produces optimal rule lists on practical problems in seconds. This framework is a novel alternative to CART and other decision tree methods.
More on http://www.kdd.org/kdd2017/
Видео Learning certifiably optimal rule lists for categorical data канала KDD2017 video
Показать
Комментарии отсутствуют
Информация о видео
Другие видео канала
![Designing AI at Scale to Power Everyday Life](https://i.ytimg.com/vi/7dn3VNkopfQ/default.jpg)
![Estimation of Recent Ancestral Origins of Individuals on a Large Scale](https://i.ytimg.com/vi/Gjy82uOnnho/default.jpg)
![struc2vec: Learning Node Representations from Structural Identity](https://i.ytimg.com/vi/lu0xMOO48Xo/default.jpg)
![Learning to Generate Rock Descriptions from Multivariate Well Logs with Hierarchical Attention](https://i.ytimg.com/vi/DXXWolWbbzY/default.jpg)
![A Local Algorithm for StructurePreserving Graph Cut](https://i.ytimg.com/vi/l5PRRHuhUQU/default.jpg)
![EmbedJoin: Efficient Edit Similarity Joins via Embeddings](https://i.ytimg.com/vi/QT_hwmQy_DU/default.jpg)
![Multi-Aspect Streaming Tensor Completion](https://i.ytimg.com/vi/PNC2_Kg8Qvo/default.jpg)
![Planning Bike Lanes based on SharingBikes' Trajectories](https://i.ytimg.com/vi/K9aUARUrt_Y/default.jpg)
![Tripoles: A New Class of Relationships in Time Series Data](https://i.ytimg.com/vi/LG1_JO-tsU4/default.jpg)
![Industrial Machine Learning](https://i.ytimg.com/vi/6u-Mz-RDUGg/default.jpg)
![Optimal Reserve Price for Online Ads Trading Based on Inventory Identification](https://i.ytimg.com/vi/JitmXoTVbjE/default.jpg)
![Multi-Aspect Streaming Tensor Completion](https://i.ytimg.com/vi/g8ZbBp8xk6U/default.jpg)
![Dispatch with Confidence: Integration of machine learning](https://i.ytimg.com/vi/lj95WCwgmoY/default.jpg)
![Interpretable Predictions of Tree-based Ensembles via Actionable Feature Tweaking](https://i.ytimg.com/vi/KIP2N5HZRW8/default.jpg)
![HinDroid: An Intelligent Android Malware Detection System](https://i.ytimg.com/vi/pQbExw-gN7k/default.jpg)
![Revisiting power-law distributions in spectra of real world networks](https://i.ytimg.com/vi/4W8lNRpcbiQ/default.jpg)
![KDD Business Lunch](https://i.ytimg.com/vi/_poHfHuTjR4/default.jpg)
![Traffic sensor health monitoring using spatiotemporal graphical modeling](https://i.ytimg.com/vi/nLxE0lbp4NQ/default.jpg)
![A Practical Algorithm for Solving the Incoherence Problem of Topic Models In Industrial Applications](https://i.ytimg.com/vi/5zoPXPbI1Jw/default.jpg)
![Detecting Network Effects: Randomizing Over Randomized Experiments](https://i.ytimg.com/vi/1v5_CzdRVAc/default.jpg)