A gradient boosting decision tree approach for insider trading identification: An empirical model evaluation of China stock market

Insider trading is a kind of criminal behavior in stock market by using nonpublic information. In recent years, it has become the major illegal activity in China’s stock market. In this study, a combination approach of GBDT (Gradient Boosting Decision Tree) and DE (Differential Evolution) is proposed to identify insider trading activities by using data of relevant indicators. First, insider trading samples occurred from year 2007 to 2017 and corresponding non-insider trading samples are collected. Next, the proposed method is trained by the GBDT, and initial parameters of the GBDT are optimized by the DE. Finally, out-of-samples are classified by the trained GBDT–DE model and its performances are evaluated. The experiment results show that our proposed method performed the best for insider trading identification under time window length of ninety days, indicating the relevant indicators under 90-days time window length are relatively more useful. Additionally, under all three time window lengths, relative importance result shows that several indicators are consistently crucial for insider trading identification. Furthermore, the proposed approach significantly outperforms other benchmark methods, demonstrating that it could be applied as an intelligent system to improve identification accuracy and efficiency for insider trading regulation in China stock market.     

Nonuniform linear array DOA estimation using EM criterion

In this letter, we address the problem of Direction of Arrival (DOA) estimation with nonuniform linear array in the context of sparse Bayesian learning (SBL) framework. The nonuniform array output is deemed as an incomplete-data observation, and a hypothetical uniform linear array output is treated as an unavailable complete-data observation. Then the Expectation-Maximization (EM) criterion is directly utilized to iteratively maximize the expected value of the complete-data log likelihood under the posterior distribution of the latent variable. The novelties of the proposed method lie in its capability of interpolating the actual received data to a virtual uniform linear array, therefore extending the achievable array aperture. Simulation results manifests the superiority of the proposed method over off-the-shelf algorithms, specially on circumstances such as low SNR, insufficient snapshots, and spatially adjacent sources.     

Probing the degenerate pattern growth of {100}<011> orientation in a directionally solidified Al-4.5 wt% Cu alloy

Degenerate pattern is a seemingly disordered morphology but it exhibits the inherently ordered crystal connected with tip-splitting and limited stability which makes it difficult to observe in the metallic system. Here we employ (100)[011] orientated planar-front seeds using directional solidification and reveal the fundamental origins of the degenerate pattern growth in an Al-4.5 wt% Cu alloy. We find that the spacing of the tip-splitting (λ) in the degenerate of the alloys followed a power law, λ∝V^−0.5, and the frequency (f) of the splitting was related to the growth velocity (V) by ƒ∝V^1.5. The dimensionless growth direction (θ/θ₀) increased monotonously and approached 0.6 with faster velocity, attributed to its anisotropy in the interface kinetics. Once growth velocity exceeded a threshold, two types of pattern transitions from degenerate to regular dendrites were proposed. One of them exhibited a random and chaotic mode and the other underwent a rotation in growth direction.     

Computational analysis of thin coating layer failure using a cohesive model and gradient plasticity

Thermal barrier coatings (TBC) are widely used to prevent transient high temperature attack and allow components high durability. Due to strong inhomogeneous material properties the TBC failure often initiates near the interface between the brittle oxide layer and the ductile substrate. A reliable prediction of the TBC failure requires detailed information about the crack tip field and the consequent fracture criteria. In the present paper both cohesive model and gradient plasticity are used to simulate the failure process and to study interdependence of the interface stress distribution with the specific fracture energies. Computations confirm that combination of the two models is able to simulate different failure mechanisms in the TBC system. The computational model has the potential to give a realistic prediction of the crack propagation process.     

由光率体求导折射率及振动方向的新方法

从光率体出发，利用坐标轴旋转，导出了光在晶体中传播时折射率及振动方向的公式。     

野外大气扩散试验中四面体气球示踪实验的初步评价

1985年11月在瑞士北部进行了一次以探讨“冷池”条件下的扩散和湍流特征为目的的国际野外大气试验,本文介绍其中的四面体气球示踪实验的结果。一共进行了三次由雷达跟踪的四面体示踪实验,释放了10个四面体球。按多轨迹法和单轨迹法分别估算并比较了其中两次示踪实验得到的水平扩散参数σ_y,探讨了风摆效应的贡献。采用“体源模式”并结合单轨迹法估算了微弱风场不定风向条件下的扩散参数,分析了静风,弱风条件下的水平风向标准差和湍流强度。     

Fast determination of critical eluent composition for polymers by gradient chromatography

Gradient chromatography was applied in order to calculate the composition at elution for different methacrylates on normal phase columns. In addition the composition at elution was determined for polyethyleneoxide on a reverse phase column. It is shown that high molar mass polymers elute for a given homopolymer irrespective of their molar mass at the same eluent composition, which varies only slightly with gradient slope. In general the composition at elution in gradient chromatography is expected to be slightly lower than the true critical composition. For high molar mass polymers we found this composition to be close to the critical composition determined by isocratic experiments. The difference between the composition at elution and the true critical composition for a variety of polymethacrylates and for polyethyleneglycol was found to be only between 0.2 and 5%. Thus, after estimating the composition at elution, only a small number of additional isocratic experiments is needed to find the exact critical composition.     

Synthesis and characterization of functional gradient copolymers of glycidyl methacrylate and butyl acrylate

Epoxy-functional spontaneous gradient copolymers of glycidyl methacrylate (G) and n-butyl acrylate (B) were synthesized via atom transfer radical polymerization (ATRP). The copolymerization reactions were carried out in toluene solution at 70 °C, using methyl 2-bromopropionate (MBrP) as initiator and copper chloride with N,N,N′,N′′,N′′-pentamethyldiethylenetriamine (PMDETA) as the catalyst system. The kinetic behaviour of the statistical copolymerizations was studied in a wide composition interval with molar fractions of G ranging from 0.10 to 0.75. The synthesized copolymers were characterized by size exclusion chromatography (SEC) and nuclear magnetic resonance (NMR) spectroscopy. ¹H NMR was employed to determine the copolymer composition, demonstrating the gradient character of the copolymers along the main chain in the whole monomer conversion interval. Apart from this, the sequence distribution and stereoregularity were analyzed. These microstructural experimental data agreed well with those calculated from Mayo-Lewis terminal model (MLTM) and a Bernoullian statistic with an isotacticity parameter of σ_G = 0.28 and a coisotacticity parameter of σ = 0.30.     

自聚焦共焦探测系统轴向分辨率的影响因素

从微小内尺度测量角度出发，研究了针孔尺寸、探测器位置、透镜的数值孔径、放大倍率和杂散光等对自聚焦共焦式探测系统轴向分辨率的影响。结果表明，为提高系统的轴向分辨率，需将有效针孔尺寸控制在≤2.5；探测器需配以精密微调整装置以实现横向精确定位；选用差动共焦光路及大数值孔径和大放大倍率的自聚焦透镜，同时，偏振光路的使用将有效的提高系统的信噪比。     

An optimization algorithm based on stochastic sensitivity analysis for noisy objective landscapes

A function minimization algorithm that updates solutions based on approximated derivative information is proposed. The algorithm generates sample points with Gaussian white noise, and approximates derivatives based on stochastic sensitivity analysis. Unlike standard trust region methods which calculate gradients with n or more sample points, where n is the number of variables, the proposed algorithm allows the number of sample points M to be less than n. Furthermore, it ignores small amounts of noise within a trust region. This paper addresses the following two questions: how does the derivative approximation become worse when the number of sample points is small? Can the algorithm find good solutions with inexact derivative information when the objective landscape is noisy? Through intensive numerical experiments using quadratic functions, the algorithm is shown to be able to approximate derivatives when M is about n/10 or more. The experiments using a formulation of the traveling salesman problem show that the algorithm can find reasonably good solutions for noisy objective landscapes with inexact derivatives information.