146 research outputs found
An emotion analysis dataset of course comment texts in massive online learning course platforms
No full textDatasets are critical for emotion analysis in the machine learning field. This study aims to explore emotion analysis datasets and related benchmarks in online learning, since, currently, there are very few studies that explore the same. We have scientifically labeled the topic and nine-category emotion of 4715 comment texts in online learning platforms using the “three-person voting label method” based on the “sentence-level” and multi-category labeling dimensions with our self-developed system. After testing the consistency of the labeling results using the Fleiss Kappa method, we found that the consistency of the dataset was about 0.51, representing a moderate strength of agreement. Based on the dataset, the prediction accuracy of the Long-Short Term Memory (LSTM) method is about 0.68. This dataset provides a benchmark for the multi- category emotion dataset in the Chinese online learning field. It can provide a basis for the subsequent solution of emotion analysis, monitoring, and intervention in the education field. It can also provide a reference for constructing subsequent datasets in the education field.We need to remind you that this is a Chinese dataset. If you want to use this dataset, please contact the author and you should request for the dataset below
SNP-based identification of QTLs for thousand-grain weight and related traits in wheat 8762/Keyi 5214 DH lines
No full textAs important yield-related traits, thousand-grain weight (TGW), grain number per spike (GNS) and grain weight per spike (GWS) are crucial components of wheat production. To dissect their underlying genetic basis, a double haploid (DH) population comprised of 198 lines derived from 8762/Keyi 5214 was constructed. We then used genechip to genotype the DH population and integrated the yield-related traits TGW, GNS and GWS for QTL mapping. Finally, we obtained a total of 18 942 polymorphic SNP markers and identified 41 crucial QTLs for these traits. Three stable QTLs for TGW were identified on chromosomes 2D (QTgw-2D.3 and QTgw-2D.4) and 6A (QTgw-6A.1), with additive alleles all from the parent 8762, explaining 4.81–18.67% of the phenotypic variations. Five stable QTLs for GNS on chromosomes 3D, 5B, 5D and 6A were identified. QGns-5D.1 was from parent 8762, while the other four QTLs were from parent Keyi 5214, explaining 5.89–7.08% of the GNS phenotypic variations. In addition, a stable GWS genetic locus QGws-4A.3 was detected from the parent 8762, which explained 6.08–6.14% of the phenotypic variations. To utilize the identified QTLs, we developed STARP markers for four important QTLs, Tgw2D.3-2, Tgw2D.4-1, Tgw6A.1 and Gns3D.1. Our results provide important basic resources and references for the identification and cloning of genes related to TGW, GNS and GWS in wheat
Essays on corporate social responsibility
No full textThe topic of this dissertation is Corporate Social Responsibility (CSR), with a focus on CSR disclosure and measurement. The dissertation consists of three chapters, including a literature review and two empirical studies on impression management in companies' CSR reports based on a dataset of CSR reports issued by U.S. companies between the year 2005 and 2018.
Chapter 1 is a literature review of CSR disclosure and measurement. The demand for reliable CSR data is rising rapidly, as investors increasingly use nonfinancial information as screening criteria when making investment decisions. However, critiques have arisen from both academia and the industry about CSR-rating products. Researchers, for example, have expressed concerns about the credibility of the measurements commonly used in research, such as KLD scores. This literature review aims to give users of CSR information an overview of publicly available data sources. It also investigates major products that provide CSR ratings for public companies and analyzes concerns expressed about those products. Finally, it summarizes the difficulties in constructing CSR performance scores and discusses the open questions in CSR disclosure and measurement and outlines potential directions for future research.
Chapter 2 and Chapter 3 test whether companies use images and linguistic features, respectively, for impression management purposes and the consequences of doing so. Chapter 2 is a study on the use of images in CSR reports. Prior research identifies images as a tool for corporate impression management. We examine (1) associations between CSR report image usage and motivations for impression management, (2) changes in CSR report image usage after controversial events, and (3) shareholder and award giver reactions to excessive image usage. We document that socially problematic industries exhibit higher image usage than others. We also document that firms who issue less extensive disclosure content or do not voluntarily commit to Global Initiative Reporting (GRI) guidelines in CSR reports tend to use more images. We find that firms with poorer ratings of CSR performance use more images in CSR reports and find some evidence indicating that firms increase image usage after controversial events. We find that excessive image usage is weakly associated with increased equity overvaluation and reduced shareholder activism, but no association with CSR awards. Overall, the evidence is consistent with companies strategically using images in CSR reports to enhance stakeholder perception of CSR engagement and performance. Such strategic usage has impacts on some audiences, though less than experimental evidence leads us to expect.
Chapter 3 is a study of the linguistic characteristics of CSR reports. Based on a classification of financial information disclosures as soft or hard, this study 1) uses linguistic features of CSR reports to proxy for hard and soft disclosure and to investigate whether companies strategically use the two types of information in CSR reports according to companies’ CSR performance, 2) examines the predictive power of the two types of information on future CSR performance, and 3) investigates the impacts of the information on decision-making. This study finds that companies adopt two strategies of communication. Those with low CSR strengths disclose less hard information, that is, numerical and specific information, and more soft “filler” language in their CSR reports. Those with high CSR concerns disclose more soft, forward-looking statements after controlling for government and media monitoring. What’s more, though forward-looking statements lack overall predictive power, those statements can better predict future performance when the CSR reports are assured. As for the consequences of the linguistic disclosure, hard information can weakly reduce shareholder activism and increase the likelihood of recognition via external awards. In contrast, boilerplate language does not benefit companies. Overall, the results suggest that the strategic use of linguistics in CSR reports does not significantly induce information users to make decisions in favor of the companies.Ph.D.Includes bibliographical reference
[[alternative]]The Pedagogical Grammar of the Chinese Modal Verb "Keyi"
No full text[[abstract]]本論文的目的在研究漢語能願動詞「可以」之教學語法。
第一章導論說明本論文研究緣起與研究目的。
第二、三章為理論分析部份，首先探討能願動詞「可以」的語義範圍與語義結構，整理出能願動詞「可以」可表達之語義有四：「表可能」、「表許可」、「表某種用途」、「表值得」，同時並與另一語義相近之能願動詞「能」做一比較，接著為「可以」整理出訂定教學語法規則時所應遵循的教學語法架構。
第四章為語料庫之分析，利用台灣師大「華語病句語料庫」之語料，整理分析華語學習者學習能願動詞「可以」各語義時的習得過程，並就學習者常犯錯誤做一分析。
第五章為教學語法部份，以上述各理論與分析為基礎與框架，嘗試提出「可以」的建議性教學排序與教學語法，並以整理之結果評比與檢視現有華語文教材對「可以」的處理方式。
第六章回顧所有論述，為總結與省思。
Chenghai hua xu zhi dai ci "yi" de yu fa fen xi.
No full text文章討論了潮汕澄海話虛指的第三人稱"伊"的語法特點并進行澄海話、普通話和粤語蒙受結構的對比分析。第一部分描寫了澄海話三種虛指代詞"伊"的相關結構──"乞伊"結構、"佮伊"結構和"V伊R"結構的語法特點。"乞伊"結構和"佮伊"結構各可分為A、B兩類，兩種結構內部分別在動詞的選擇、"伊"的所指和"乞/佮"的隱現上表現出不同。兩種結構之間的主要差別則在於主語的題元角色及"乞/佮"的句法位置高低上。"V伊R"結構是"佮伊"結構的一種條件變體，主要用於表達祈使等語氣。文章第二部分分析了澄海話三種蒙受結構的生成機制。并進一步概括出"伊"的種類:從所指對象題元角色的角度可以分為兩類，一是施事或致事、二是蒙事；從所指對象位置的角度也可以分為兩類，一是所指在句中，二是所指在句外。以及"伊"的句法位置：一是，位於"中間層次"；二是，緊鄰輕動詞。文章第三部分嘗試以澄海話蒙受結構的分析方法為角度和基礎，分別分析普通話蒙受結構"V他O"和粤語蒙受結構"VR佢"。通過三種方言蒙受結構的對比分析，我們看到三種方言類型上的不同，也驗證了劉丹青（2001）、鄧思穎（2006）的分析：澄海閩南話是"最弱的SVO"，動詞傾向於留在原地；粤語是"最強的SVO"，動詞傾向於長距離移位；普通話是"溫和的SVO"，動詞移位但距離不長。This thesis discusses the syntax of the Chinese non-referential pronoun 伊 yi ‘it’ in Chenghai Southern Min in three different affective constructions: keyi 乞伊 construction, kaiyi 佮伊 construction and V伊R construction. After examining the syntactic and semantic properties of these three yi affective constructions, we figure out that firstly there are two types of keyi 乞伊 construction and two types kaiyi 佮伊 construction. Secondly keyi 乞伊 construction differs from kaiyi 佮伊 construction in theta-roles of the subject and the position of yi. Thirdly V伊R construction is a conditioned variant of keyi 乞伊 construction. The following part we explore the hierarchical order of these three yi affective constructions, and we employ high and middle applicatives to capture their syntactic and semantic idiosyncrasies. In this part yi is classified into two ways according to the property of the object it refers to. For example, in keyi 乞伊 construction, yi always refers to a causer or a agent, and in kaiyi 佮伊 construction, yi refers to an affectee. In both constructions, yi is situated in a middle position which close to light verbs. The last part of this paper compare the affective constructions of Chenghai Southern Min with affective constructions of Mandarin and Cantonese. The main differences among these three dialects match with the claim proposed by Liu(2001) and Tang(2006): Chenghai Southern Min is the weakest SVO construction in which verbs prefer to stay in situ; Cantonese is the strongest SVO construction in which verbs prefer to take long-distance movement; Mandarin is the moderate SVO construction in which verbs will move but won’t move far.Detailed summary in vernacular field only.Detailed summary in vernacular field only.Detailed summary in vernacular field only.金佳.Thesis (M.Phil.) Chinese University of Hong Kong, 2014.Includes bibliographical references (leaves 87-93).Abstracts in English and Chinese.Jin Jia
Resource-Transition Circuits and Siphons for Deadlock Control of Automated Manufacturing Systems
No full textThe resource-transition circuit ( RTC) and siphon are two different structural objects of Petri nets and used to develop deadlock control policies for automated manufacturing systems. They are related to the liveness property of Petri net models and thus used to characterize and avoid deadlocks. Based on them, there are two kinds of methods for developing deadlock controllers. Such methods rely on the computation of all maximal perfect RTCs and strict minimal siphons (SMSs), respectively. This paper concentrates on a class of Petri nets called a system of simple sequential processes with resources, establishes the relation between two kinds of control methods, and identifies maximal perfect RTCs and SMSs. A graph-based technique is used to find all elementary RTC structures. They are then used to derive all RTCs. Next, an iterative method is developed to recursively construct all maximal perfect RTCs from elementary ones. Finally, a one-to-one correspondence between SMSs and maximal perfect RTCs and, hence, an equivalence between two deadlock control methods are established. © 2006 IEEE
Simulation of Lightning Protection for Composite Civil Aircrafts
No full textAbstractLightning damage will be a hazard to composite aircrafts if they are not properly designed. In this paper, based the aircraft lightning protection theory, lightning zoning and its protection technology for a certain type of composite civil aircraft are studied utilizing EMA3D electromagnetic simulation software, which can provide relevant information for design and verification of lightning protection for civil aircrafts
Maneuvering Target Tracking in Dense Clutter Based on Particle Filtering
No full textAbstractAn improved particle filtering (IPF) is presented to perform maneuvering target tracking in dense clutter. The proposed filter uses several efficient variance reduction methods to combat particle degeneracy, low mode prior probabilities and measurement-origin uncertainty. Within the framework of a hybrid state estimation, each particle samples a discrete mode from its posterior distribution and the continuous state variables are approximated by a multivariate Gaussian mixture that is updated by an unscented Kalman filtering (UKF). The uncertainty of measurement origin is solved by Monte Carlo probabilistic data association method where the distribution of interest is approximated by particle filtering and UKF. Correct data association and precise behavior mode detection are successfully achieved by the proposed method in the environment with heavy clutter and very low mode prior probability. The performance of the proposed filter is examined and compared by Monte Carlo simulation over typical target scenario for various clutter densities. The simulation results show the effectiveness of the proposed filter
The Impact of Caregiver's Usage of Internet on Teenager's Cognitive Skills
No full textIn the digital age, internet becomes an important source of information gathering and maintains network of human social contact. In this paper, we use micro data from CFPS to examines the effects of guardian's choice of using internet on teenager's cognitive skills. Using fixed effect model and instrumental variable model, we find that guardian's internet usage has a sizable and significant positive effect on teenager's cognitive ability measured by math and verbal test scores. Further exploration of mechanism shows that internet usage affects cognitive skills through direct channels (increasing learning resources) and indirect channels (improving parenting skills). Heterogeneous analysis shows that the impact is stronger for guardians in rural areas and less-educated guardians. The result suggests that promoting internet for the less educated family in rural areas can potentially improve children's academic performance and decrease inequality transmitted across generation.隨著數字經濟的發展，互聯網成为重要的收集信息和構建社交網絡的渠道。在本文中，我們使用CFPS的微觀數據來檢驗監護人使用互聯網的情况對青少年認知的影響。使用基站密度和移動互聯網用户數量作为工具變量，我們的研究結果表明，監護人的互聯網使用對青少年的認知能力有顯著的積極影響。文中的認知能力通過數學和語言測試來衡量。對機制的進一步探索表明，監護人使用互聯網，增加了孩子教育投資，改善了監護人的教育理念，進而來影響青少年認知。异質性分析表明，農村地區的監護人或受教育水平較低的監護人受到的影響更大。結果表明，鼓勵農村地區受教育程度較低的家庭使用互聯網可以潜在地提高孩子的學業表現進而减少代際傳遞的不平等。GUO, Keyi.M.Phil. Chinese University of Hong Kong 2021.Includes bibliographical references (leaves )Abstracts in English and Chinese.Title from PDF title page (viewed on ...
- …
