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.

Realistic simulation of plant based on geometry parameter

Abstract: Plants play an important role in the natural scenery of landscape. Unfortunately, due to the high level complexity of the structure of plants, simulating tree becomes extremely difficult. The paper used fractal recursion algorithm to construct the 3D tree model whose parameters were variable. We took advantage of texture mapping, simulated shadow and made use of Bezier surface to enhance the sense of reality of the project. A test example of blooming in a tree created solely by OpenGL graphics library was provided to show the feasibility of the algorithm.

Decision tree construction method based on rough set and distance function

Abstract: Decision tree which is effectively used in the classification of data mining.On the decision tree construction algorithm,the core of condition attributes with respect to decision attributes in rough set theory is used for selection of attributes in multivariate tests.Considering the advantage and disadvantage of the decision trees and rough set,the decision trees and rough set are combined.A new multivariate decision tree construction algorithm is offered,which restricts each of nodes containing the number of attributes,and then,choice attributes combination according to the attribute dependability and De Mantaras distance function.The prominent merit of this al-gorithm is that it can reduce the height of a tree,and can raise readability of classing rules.

A Fast-Automated Scheduling Method Based on Genetic Algorithm for Final Assembly Workshop

Abstract: In this paper, we proposed a fast-automated scheduling method for the final assembly workshop based on the Genetic Algorithm. This method belongs to the technical field of production planning management involving the application of the genetic algorithm and computer simulation in advanced planning and scheduling. Comprehensively, we take into account realistic factors including order priority, lines workload balance and time-consuming to design our three-level scheme evaluation systems. Moreover, the order priority is decided by the Decision Tree model we train. Considering the efficiency, we redesign the coding scheme, variation operator and fitness function to make the Genetic Algorithm serve as an exclusive accelerator for the approximation of optimal schemes. As a result, this fast-automated scheduling method is practical and conducive to production efficiency improvement as well as enterprise benefits maximization.

Predicting Event Time by Classifying Sub-Level Temporal Relations Induced from a Unified Representation of Time Anchors.

Abstract: Extracting event time from news articles is a challenging but attractive task. In contrast to the most existing pair-wised temporal link annotation, Reimers et al.(2016) proposed to annotate the time anchor (a.k.a. the exact time) of each event. Their work represents time anchors with discrete representations of Single-Day/Multi-Day and Certain/Uncertain. This increases the complexity of modeling the temporal relations between two time anchors, which cannot be categorized into the relations of Allen's interval algebra (Allen, 1990). In this paper, we propose an effective method to decompose such complex temporal relations into sub-level relations by introducing a unified quadruple representation for both Single-Day/Multi-Day and Certain/Uncertain time anchors. The temporal relation classifiers are trained in a multi-label classification manner. The system structure of our approach is much simpler than the existing decision tree model (Reimers et al., 2018), which is composed by a dozen of node classifiers. Another contribution of this work is to construct a larger event time corpus (256 news documents) with a reasonable Inter-Annotator Agreement (IAA), for the purpose of overcoming the data shortage of the existing event time corpus (36 news documents). The empirical results show our approach outperforms the state-of-the-art decision tree model and the increase of data size obtained a significant improvement of performance.

Reducing algorithm complexity in trees

Abstract: FSPMs make intensive use of algorithms that manipulate the branching structure of plants. These algorithms may serve a variety of computational purposes related to these branching structures such as displaying them on screen, extracting data samples and statistics from them, comparing them, computing flows of matter (water, nutrients, signals) through them, computing the propagation of physical forces in them, or growing them. In general, the computational complexity of all these algorithms is directly proportional to a power (greater than 1) of the number N of components of the considered branching structure. In most FSPMs, N is relatively high and tends to grow exponentially with the plant’s age, at least in the first stages of plant development. For this reason, most FSPM algorithms are computationally expensive, with a time and memory complexity at least proportional to N. Different strategies have been considered to reduce this computational complexity and they most of them make use of a change of scale in the plant representation. Rather than computing light interception on each leaf of a tree for example, more efficient algorithms can be achieved by assessing the light intercepted by crownlets at a coarser scale. These approaches are computationally efficient but require both a change of plant representation and of the associated models. In this paper, we present a new way of reducing algorithm complexity in FSPMs by compressing plant branching structures representation at a given scale. For this, we make use of the possibility to compress tree branching structures as directed acyclic graphs (DAGs) either in a lossless or approximate manner [1]. We show how these DAG-transforms of branching systems can be used to reformulate a wide class of algorithms that operate on trees, and that, if N denotes the number of tree components and D the number of DAG components corresponding to the compressed tree, the algorithm complexity are in general reduced from O(N) to O(D) in both time and space. For a special class of trees, called self-nested trees, with high compression capacity, the complexity is reduced from O(N) to O(Log(N)). Together with the possibility to approximate any tree by a tree in the class of self-nested trees [2], this opens up the way to perform a wide range of FSPM algorithms on any tree with remarkable spatial and temporal efficiency. We illustrate our approach on two markedly different types of algorithms for FSPMs. First, we show how the geometry of tree branching structures can be efficiently compressed and used. We then demonstrate the benefits of such an approach for the encoding of the geometric representation of various types of classical plant architectures. We then show how our approach can be used to achieve drastic simplification in the computation of flows in branching structures. Based on previous works that presented how flows can be recursively computed in trees [3], we show that compression techniques can be used to reduce substantially the complexity of flow computation in trees. We then use this new computational scheme to explore the impact of tree architecture on flow propagation in trees.