116 research outputs found
Data Mining-Driver's Injury and Passenger Characteristics
No full textThis project is related to relationship between river’s Injury and Passenger Characteristics through analyzing on Python-Pandas and Plotly. The coding is stored in file "Lingyi Meng.ipynb" and diagrams created by coding is kept as png form. There are totally two main part of datasets: "2016OhioCrash.xlsx" is the core data for analyzing and the left mini table are designed for analyzing the overlook together
Disordered packing density of binary and polydisperse mixtures of curved spherocylinders
No full textParticle elongation is an important factor affecting the packing properties of rod-like particles. However, rod-like particles can be easily bent into non-convex shapes, in which the effect of bending should also be of concerned. To explore the shape effects of elongation and bending, together with the size and volume fraction effects on the disordered packing density of mixtures of non-convex, particles, binary and polydisperse mixtures of curved spherocylinders are simulated employing sphere assembly models and the relaxation algorithm in the present work. For binary packings with the same volume, curves of the packing density versus volume fraction have good linearity, while densities are plotted as a series of equidistant curves under the condition of the same shape. The independence of size and shape effects on the packing density is verified for mixtures of curved spherocylinders. The explicit formula used to predict the density of binary mixtures, by superposing the two independent functions of the size and shape parameters, is extended to include a non-convex shape factor. A polydisperse packing with the shape factor following a uniform distribution under the condition of the same volume is equivalent to a binary mixture with certain components. The packing density is thus predicted as the mean of maximum and minimum densities employing a weighing method. (C) 2016 Chinese Society of Particuology and Institute of Process Engineering, Chinese Academy of Sciences. Published by Elsevier B.V. All rights reserved.National Natural Science Foundation of China [11272010, 11572004, 11602088]; China Postdoctoral Science Foundation [2016M592484]SCI(E)ARTICLE73-813
System development and clinical applications of the handheld PA/US dual mode imaging system
No full textAs an emerging technology, photoacoustic (PA) imaging (PAI) has gained lots of progresses in the past two decades. By combing optical contrast with ultrasound detection, PAI can maintain high spatial resolution in deeper regions. Furthermore, with the aid of multispectral imaging, PAI can reveal functional information such as oxygenation saturation (SO2), which is closely related to tumor malignancy. With the above advantages, nowadays PAI has gained lots of progresses in clinically translatable research. These research shows that PAI has the potential to complement existing imaging techniques such as ultrasound (US) imaging for cancer detection, disease evaluation and prognosis monitoring.
This thesis mainly discusses the clinical application of PAI, including system design and clinical research. Particularly, it presents several studies including the development of PA/US imaging system and the clinical evaluation of PA/US imaging in diagnosis of superficial cancer. The main research results and the major novelties of this thesis include:
1) The collaborative development of the PA/US imaging system for clinical studies, where the author was mainly responsible for the design of optical path and the trigger method between the laser and the US system.
2) The design of optical fluence compensation strategy based on clinical US structural imaging. This strategy first identifies the tissue type based on US structural imaging, and then utilizes the known optical absorption and scattering parameters to simulate optical fluence map in the imaging region. The simulated optical fluence map can then be used for correction of original PA images. The effectiveness of this strategy has been verified based on clinical PAI data. The proposed method can improve the accuracy of quantification PAI and restore PA signals in deeper regions.
3) The collaborative study of PA/US imaging on thyroid nodules with the self-developed 2D PA/US handheld imaging system. By comparing PAI results with color Doppler flow imaging (CDFI) results from 10 thyroid nodules, we found that PAI can reveal more abundant vessels than CDFI, and can thus provide valuable information in diagnosis of thyroid cancer.
4) The design of the method for quantification analysis of 3D PA/US imaging on breast cancer. Specifically, this method first calculates the ellipse enclosing the tumor region with the minimum volume and then automatically segments the tumor regions and tumor surrounding regions. Next, the volumetric mean SO2 of tumor regions and tumor surrounding regions were calculated for quantification analysis. Our results demonstrate that quantification analysis of 3D functional PA/US imaging on breast cancers has the potential to improve the specificity in diagnosis of breast cancer.
In Chapter 1, the mechanism of PAI and its recent clinical translational progresses are first reviewed. Next, in Chapter 2, the development of a clinical PA/US imaging system based on handheld US imaging is presented. The imaging ability of the newly developed system was verified with the phantom study as well as the in vivo study. After then, in Chapter 3, the optical fluence compensation method for handheld PA/US imaging based on tissue structural information provided in US imaging is introduced. The clinical result demonstrates that important tissue structures such as vessels in deeper tissue can be restored after fluence compensation based on the proposed method. Chapter 4 and Chapter 5 introduce the clinical studies based on the self-developed PA/US imaging systems, including 2D PA/US imaging of thyroid nodules, 3D PA/US imaging of breast cancer, and quantification analysis of the imaging results. The quantification analysis demonstrates that 3D functional PA/US imaging has the potential to improve the diagnosis specificity of breast cancer.
In Chapter 6, the author presented the research conducted in Georgia Institute of Technology. In this chapter, the laser induced surface acoustic waves (SAWs) and acoustic radiation force (ARF) induced SAWs were compared. The results demonstrate that laser-induced SAW imaging is able to perform stiffness evaluation and has the potential to provide higher spatial resolution.
In Chapter 7, the author summarized the major novelties in this thesis and discussed research directions in the future work.Ph.D
Maximally dense random packings of intersecting spherocylinders with central symmetry
No full textThe packing of rod-like particles has arisen in a variety of scientific and industrial applications. For the factors that attribute to the packing properties of such particles, elongation effect is one of the most important. However, rodlike particles can be easily assembled into non-convex shapes, in which the effects of non-convex deformations should be concerned. In this paper, the dense random packings of identical intersecting spherocylinders with central symmetry are numerically simulated through an analytical model and the relaxation algorithm. The maximally dense random packing (MDRP) states of 2D and 3D intersecting spherocylinders with various aspect ratios are determined from the order maps. In the MDRP states, the specific volume V, defined as the reciprocal of the packing density do, shows a highly linear correlation with the aspect ratio (w >= 1.0), which is similar to the monophasic packing of spherocylinders. This indicates that the elongation effect is the main shape factor that attributes to the packing density. Consequently, the explicit formulas to predict the packing densities of 2D and 3D intersecting spherocylinders are built as single-variable functions of the aspect ratio. The dense random packing density of 2D intersecting spherocylinders equals to that of identical spherocylinders when w=1.25. This suggests that a balance exists between the relative excluded volume effect and the multi-point contact effect to the packing density of non-convex particles. (C) 2016 Elsevier B.V. All rights reserved.National Natural Science Foundation of China [11272010, 11572004]; China Postdoctoral Science Foundation [2016M592484]CPCI-S(ISTP),SI49-5831
Military Expense Analysis
No full text<p>This project collects the data related to military expense, GNP and GDP from different countries to do the data analysis and tendency predication. The main technology utilized in project is Python-Panda</p>
Tableau Athletic Analysis
No full textThis is the sample design for international team players for school. The main part includes three sections: location, gender and degree
On the Combination of Adjective and Noun Predication : The Case where the Subject Indicates a Human
No full textTheoretical development and empirical evidence on misalignment between investor investment horizon and executive pay duration
No full textThis dissertation focuses on the issue of misalignment between executive pay duration and investor investment horizon.
In Chapter 1, this dissertation works within the agency paradigm with two alternations: First, unlike the traditional agency model, where the shareholders are assumed to be of one kind and prefer long-term value creation, we assume heterogenous time preferences of the investors (short-term investors pursuing short-term trading gains and long-term investors who pursue long-term value creation). We develop a simple three-period model (original period, period 1 and period 2), where there are two types of principals liquidate their holdings at different periods. Second, unlike the standard agency models, in which the management can affect the stock price only through the choice of effort (Spence and Zeckhauser, 1971; Ross, 1973; Holmstrom, 1979), our model assumes the manager can engage in unobservable earnings manipulation, like in Goldman and Slezak (2006), that upwardly biases disclosed information but hurts the long-term cash flows of the firm. The manager chooses a myopic action at the original period that can increase the period 1(short-term) earnings at the expense of the period 2 (long-term) value. Therefore, the unobservable earnings manipulation transfers long-term value of the firm to the current. We first work out the manager's hidden choice that maximizes the manager's payoff based on the contract that has been offered prior to the choice of the action and ownership structure. Then we take a step further to examine the optimal contract scheme by making the contract endogenously set by the current shareholders and argue that the optimal contract should “mirror” the ownership structure of the firm.
In Chapter 2, we use empirical method to test the prediction from our model in Chapter 1, that short-termism is a result of optimal compensation contracting, by showing a positive relationship between investors investment horizon, as measured by the reverse of the weighted churn rate of firms’ institutional investors, and CEO’s pay duration, as measured by the weighted average of different components of CEO’s compensation grants, after controlling other factors may affect pay duration industry fixed effect and year effect. Then, we take a further step to develop a measure to quantify the sub-optimality of the compensation contract by measuring the misalignment between investor horizon and executive pay duration. We take the absolute value of the difference between standardized investor investment horizon and standardized CEO pay duration and identify that the misaligned horizons will lead to higher stock volatility and more harmful real earnings management (negative effect on future operating performance). The negative result of misalignment is more pronounced if the pay duration is shorter than the investor horizon. That is to say, the abnormal behavior (overproduction, sales discount, extraordinary expense cut) captured by Roychowdhury's model is more out of opportunistic choice rather than efficient contract when their pay duration is misaligned with investors' investment horizon, especially with the misalignment direction of shorter pay duration relative to investor horizon. Therefore, we argue that what really matters to existing shareholders’ interest is not whether the compensation scheme of the management is long-term or short-term per se, but rather the relative stand of executive pay duration comparing with their own investment horizon.Ph.D.Includes bibliographical referencesby Lingyi Zhen
- …
