Decision tree model

About: Decision tree model is a research topic. Over the lifetime, 2256 publications have been published within this topic receiving 38142 citations.

Decision Tree Visualization for High-Dimensional Numerical Data

Abstract: Big Data enforces the usage of data mining techniques to provide the user valuable insights. There is a broad range of data mining and machine learning techniques tackling different tasks. Generic approaches are classification algorithms, which label given data points by a pretrained model. Decision tree-based classification algorithms are often used, as they provide a human-explainable model, which can be represented by simple induced rules. In order to present the classification results and the concrete model to the user, there exist for both problems a set of different solutions. Current visualizations either project labeled data into the plane or three-dimensional space, or the visualizations illustrate the decision tree rules as e.g. graph structures. But they lack to provide a possibility to show both, data and the model, within a single plot. Therefore, we propose a projection strategy to present both decision tree model and data in a single plot. Furthermore, we developed an interactive visualization to showcase the proposed approach and evaluated the visualization on open-source datasets. The results show that the plots can be computed in short time and projection adjustments are reasonably low.

On Network Performance Indicators for Network Promoter Score Estimation

Abstract: Estimation of user perceived quality of offered services, from massive number of Key Performance Indicator (KPI)’s that are measured in diverse components, has been a necessity for mobile network operators. The goal is first to have a good estimator for poor Quality of Experience (QoE), which can potentially be achieved with machine learning, and then pinpoint the features that are contributing to the poor performance. There is often a tradeoff between accuracy and interpretability of models. In this paper, we address this tradeoff by first developing a robust but complex teacher machine learning model to map the subjective Net Promoter Score (NPS) values computed from the user quality feedback to the underlying subset of KPI metrics. Next, we develop a rather interpretable student model supervised by the pre-trained teacher model. Eventually the compact student decision tree model learns to mimic the behavior of the teacher model with an at least 10 % improved accuracy in testset as compared to conventional way of directly training using the decision tree model. In the last step, we extract the rules and important influential features of the distilled student model.

Computational Complexity of Tree Evaluation Problems and Branching Program Satisfiability Problems

Abstract: Computational Complexity of Tree Evaluation Problems and Branching Program Satisfiability Problems