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.

Copula based hierarchical risk aggregation - Tree dependent sampling and the space of mild tree dependence

Abstract: The ability to adequately model risks is crucial for insurance companies. The method of "Copula-based hierarchical risk aggregation" by Arbenz et al. offers a flexible way in doing so and has attracted much attention recently. We briefly introduce the aggregation tree model as well as the sampling algorithm proposed by they authors. An important characteristic of the model is that the joint distribution of all risk is not fully specified unless an additional assumption (known as "conditional independence assumption") is added. We show that there is numerical evidence that the sampling algorithm yields an approximation of the distribution uniquely specified by the conditional independence assumption. We propose a modified algorithm and provide a proof that under certain conditions the said distribution is indeed approximated by our algorithm. We further determine the space of feasible distributions for a given aggregation tree model in case we drop the conditional independence assumption. We study the impact of the input parameters and the tree structure, which allows conclusions of the way the aggregation tree should be designed.

HPFLRF: A High Performance Fingerprint Localization Algorithm Based on Random Forest

Abstract: In recent years, indoor localization base on fingerprint has become more common. Due to the complexity and variability of indoor environment, it is difficult for traditional indoor localization algorithm to obtain better localization accuracy and stability. In this paper, we propose a high performance fingerprint localization algorithm based on random forest (HPFLRF), which has higher precision and stability. Our algorithm could select a valid subset of APs through multiple AP selection method. In addition, our algorithm uses the random forest training positioning model to improve the stability of the algorithm effectively, and overcome the problem of overfitting in single decision tree model. The results of experiment show that our algorithm has better localization performance which average positioning error is 1.3718 m, only one seventh of the localization algorithms based on multiple times AP selection and decision tree.

Cross-domain computing task scheduling method and system based on intelligent perception

Abstract: The invention provides a cross-domain computing task scheduling method and system based on intelligent perception. The cross-domain computing task scheduling method comprises the steps of step 1, training a decision tree model based on label data; step 2, estimating and calculating the execution time of the task based on the relative time complexity; step 3, predicting a resource change trend index of each domain based on the resource historical record and an ARIMA algorithm; step 4, obtaining resource real-time state indexes of each domain by using a resource state interface; step 5, estimating migration time of the data migrated to each domain based on the available bandwidth; and step 6, deciding the task optimal execution domain based on the decision tree model and the comprehensive index. The trend prediction algorithm and the decision tree algorithm are creatively and comprehensively applied to the cross-domain computing task scheduling scene, the task resource preemption phenomenon is avoided, and the problem that the scheduling decision accuracy is low is solved; through the flow type machine learning technology, the performance problems of a trend prediction algorithm anda decision tree algorithm are solved, and the overall time of cross-domain computing task scheduling is greatly shortened.

The Use of Decision Tree Analysis to Improve LDA Decision Making

Abstract: In food animal practice, medical decisions are usually made on the basis of economic impact to the dairy, rather than on perceived individual animal value or emotional attachment. A decision routinely made on dairies is what to do with cows that develop left displaced abomasums (LDA). Given the variety of management options available, veterinarians have the potential to make different recommendations depending upon the cow's historical value, parity, current stage oflactation, presence of concurrent disease, level of milk price and relative replacement cost. Decision trees are systematic quantitative tools that may be used to improve the ability to select the best course of action in situations, such as LDA, where the clinical decision is complex and outcomes are uncertain. The objective of this project was to demonstrate the use of a decision tree model as a tool to help select between surgery, roll and toggle or marketing for beef as the most economically appropriate management plan, given an early lactation cow with an LDA.

On average depth of decision trees implementing Boolean functions

Abstract: The article considers the representation of Boolean functions in the form of decision trees. It presents the bounds on average time complexity of decision trees for all classes of Boolean functions that are closed over substitution, and the insertion and deletion of unessential variables (the structure of these classes is described in the book by Jablonsky, Gavrilov and Kudriavtzev [5]. The obtained results are compared with the results developed by Moshkov in [6] that describe the worst case time complexity of decision trees.