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.

Probabilistic Label Trees for Efficient Large Scale Image Classification

Abstract: Large-scale recognition problems with thousands of classes pose a particular challenge because applying the classifier requires more computation as the number of classes grows. The label tree model integrates classification with the traversal of the tree so that complexity grows logarithmically. In this paper, we show how the parameters of the label tree can be found using maximum likelihood estimation. This new probabilistic learning technique produces a label tree with significantly improved recognition accuracy.

Decision tree design from a communication theory standpoint

Abstract: A communication theory approach to decision tree design based on a top-town mutual information algorithm is presented. It is shown that this algorithm is equivalent to a form of Shannon-Fano prefix coding, and several fundamental bounds relating decision-tree parameters are derived. The bounds are used in conjunction with a rate-distortion interpretation of tree design to explain several phenomena previously observed in practical decision-tree design. A termination rule for the algorithm called the delta-entropy rule is proposed that improves its robustness in the presence of noise. Simulation results are presented, showing that the tree classifiers derived by the algorithm compare favourably to the single nearest neighbour classifier. >

Tree modelling from mobile laser scanning data-sets

Abstract: In recent times mobile laser scanning (MLS) has been used to acquire massive 3D point clouds in urban areas and along road corridors for the collection of detailed data for 3D city modelling, building facade reconstruction and capture of vegetation and road features for inventories. The objectives of this paper are the extraction of tree features from such data-sets and the modelling of trees for the purpose of visualisation in 3D city models. After the detection of high vegetation the point cloud is reduced using a 3D alpha shape approach. Then the required model parameters such as crown and stem height, crown and stem diameter, and crown shape are derived and the trees are modelled individually in a realistic manner. The tree model so generated correctly represents the overall appearance of the tree. However, the inner structure such as the branching of the tree crown is parameterised. The workflow reduces the point cloud by means of a step-by-step process, which eases the handling of the massive MLS data-sets. The thinning using 3D alpha shapes reduces the amount of data to be processed by about 95%. It is shown that the model parameters are not influenced by the thinning procedure employed. This proves the robustness of the data reduction method and the tree modelling approach.

The sketch L-system : Global control of tree modeling using free-form strokes

Abstract: L-system is a tool commonly used for modeling and simulating the growth of plants. In this paper, we propose a new tree modeling system based on L-system that allows the user to control the overall appearance and the depth of recursion, which represents the level of growth, easily and directly, by drawing a single stroke. We introduce a new module into L-system whose growth direction is determined by a user-drawn stroke. As the user draws the stroke, the system gradually advances the growth simulation and creates a tree model along the stroke. Our technique is the first attempt to control the growth of a simulation in L-system using stroke input.

Time complexity of decision trees

Abstract: The research monograph is devoted to the study of bounds on time complexity in the worst case of decision trees and algorithms for decision tree construction. The monograph is organized in four parts. In the first part (Sects. 1 and 2) results of the monograph are discussed in context of rough set theory and decision tree theory. In the second part (Sect. 3) some tools for decision tree investigation based on the notion of decision table are described. In the third part (Sects. 4–6) general results about time complexity of decision trees over arbitrary (finite and infinite) information systems are considered. The fourth part (Sects. 7–11) contains a collection of mathematical results on decision trees in areas of rough set theory and decision tree theory applications such as discrete optimization, analysis of acyclic programs, pattern recognition, fault diagnosis and probabilistic reasoning.