On Binary Classification in Extreme Regions
Authors: Hamid JALALZAI, Stephan Clémençon, Anne Sabourin
NeurIPS 2018 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | Beyond theoretical results, numerical experiments are presented in order to illustrate the relevance of the approach developed. ... Numerical experiments are presented in order to illustrate the relevance of the approach developed. ... The purpose of our experiments is to provide insights into the performance of the classifier bgk on extreme regions constructed via Algorithm 1.The training set is ordered as in Step 1 of Algorithm 1. For a chosen k, let t = ˆT(Xtrain (k) ) , the L1 norm is used throughout our experiments. The extreme test set T is the subset of test points such that ˆT(Xtest i ) > t. |
| Researcher Affiliation | Academia | Hamid Jalalzai, Stephan Cl emenc on and Anne Sabourin LTCI Telecom Paris Tech, Universit e Paris-Saclay 75013, Paris, France first.last@telecom-paristech.fr |
| Pseudocode | Yes | Algorithm 1 (ERM in the extremes) |
| Open Source Code | No | The paper does not provide any explicit statements or links indicating that its source code is open or publicly available. |
| Open Datasets | Yes | The simulated dataset is generated from a logistic distribution as described in [13]. The real dataset known as Ecoli dataset, introduced in [9], deals with protein localization and contains 336 instances and 8 features. |
| Dataset Splits | No | For the experiment on the Ecoli dataset (Figure 5), one third of the dataset is used as a test set and the rest corresponds to the train set. The paper specifies train and test splits but does not mention a separate validation set split. |
| Hardware Specification | No | The paper does not provide any specific details about the hardware (e.g., GPU/CPU models, memory) used to run the experiments. |
| Software Dependencies | No | The paper mentions using 'random forest (RF)' and 'k-nearest neighbors (k-NN)' algorithms but does not specify any software libraries or their version numbers (e.g., scikit-learn, PyTorch versions) used for implementation. |
| Experiment Setup | Yes | The number of trees for both random forests (in the regular setting and in the setting of Algorithm 1) is set to 200. The number of neighbors for both k-NN s is set to 5. ... We report the results obtained with 5 103 points for each label for the train set and 5 104 points for each label for the test set. k = 100 and κ [1, 0.3]. |