7627 research outputs found
Sort by
Machine learning research on time series data
No full textTime series generated by complex systems, such as industrial IoT and user behavior systems, confront two core challenges: structured missingness (e.g., continuous or periodic gaps) that disrupt temporal dependencies, and the difficulty in effectively modeling dynamic long- and short-term temporal dependencies inherent in evolving patterns (e.g., user interests). Traditional approaches struggle to balance the preservation of local dependency continuity and the rational association of global long-range dependencies in structured missing scenarios, often incurring high computational costs. In temporal pattern modeling, the lack of adaptive mechanisms to fuse evolving long- and recent behavior trends (e.g., stable interest inertia vs. short-term preference shifts) leads to suboptimal modeling of nonlinear associations in dynamic systems.
To address these issues, this thesis conducts research from two independent dimensions—data quality enhancement and efficient temporal modeling—and constructs a dual-module methodological framework:
Structured Missingness Imputation for Time Series: A local-global dynamic fusion framework is proposed. It employs near-neighbor temporal averaging for preprocessing to retain local dependencies at missing segment boundaries, combined with a lightweight dynamic weight adjustment mechanism to adaptively integrate local interpolation and global dependency modeling. This resolves the failure of dependency reconstruction in structured missing scenarios while reducing computational overhead. Multidimensional temporal feature encoding is introduced to enhance hierarchical modeling of high-frequency fluctuations (e.g., equipment transient anomalies) and low-frequency trends (e.g., long-term operational patterns).
Dynamic Temporal Dependency-Driven Interest Transfer Modeling: An adaptive decoupling framework for long- and recent behavior trends is designed. It utilizes attentionenhanced recurrent neural networks to capture stable historical behavior trends, incorporates bidirectional temporal encoding to model forward-backward dependencies in short-term behaviors, and employs a semantic similarity gating mechanism to allocate weights between long- and recent behavior trends dynamically. This overcomes the bottleneck of traditional models in modeling nonlinear interest evolution. A domain-knowledge-driven interest quantification method is integrated to transform prior information, such as user behavior frequency and duration, into temporal interest representations, improving the model’s responsiveness to dynamic scenarios.
The proposed dual-module framework focuses on the essential characteristics of time series in complex systems, addressing issues encountered in real-world scenarios through methods such as dynamic weight adjustment and attention mechanism optimization. The research outcomes provide reusable modular solutions for industrial data governance, personalized recommendation, and other domains, advancing the analysis of complex system time series from application-specific modeling toward a methodology-driven universal technical system. Future work may explore intelligent collaboration mechanisms between the two modules and extensions to cross-modal complex systems
Two essays on health and health behavior in China
No full textThis thesis consists of two empirical studies on elderly mental health and on health behaviors. The first study investigates the long-term impact of childhood hunger on mental health in later life, using data from two nationally representative surveys in China. The study finds that childhood hunger exposure significantly increases the risks of depression and anxiety in later life. These effects are most pronounced among individuals exposed after age 5, suggesting that autobiographical memory formation plays an important role beyond socioeconomic and health pathways. The findings further indicate these impacts can be mitigated through economic interventions, such as welfare programs, and psychosocial support, including co-residing with children. The second study evaluates the behavioral and clinical effects of health information provision in China, where health screening is lacking. This study uses data from the China Health and Retirement Longitudinal Study (CHARLS), which provides respondents with blood pressure test results in each survey round. Leveraging a regression discontinuity design around different clinical hypertension thresholds, we demonstrate that individuals who receive blood pressure screening information: (1) initiate preventive behaviors when reaching pre-hypertension levels, and (2) show significant blood pressure reductions at follow-up. These effects are particularly pronounced among rural populations and less-educated individuals. Our findings highlight how routine screening programs could improve early detection and reduce cardiovascular disease burden in China\u27s primary healthcare system, with important implications for addressing health disparities
Can mere possession of a consumer product boast your self-efficacy? A cultural perspective
No full textBeggan (1992) introduced the mere ownership effect, suggesting that people evaluate owned objects more positively. Beggan argued that ownership fosters a psychological bond between people and their possessions, driven by self-enhancement motives (e.g., “I am good, thus my object must be good”). Yeung et al. (2017, 2020) extended the mere ownership effect, showing that ownership influences self-evaluations (e.g., I own a book so I am knowledgeable). Previous studies on the “person-based” mere ownership effect have primarily focused on East Asian participants. This thesis aims to compare cross-cultural differences, if any, in the “person-based” mere ownership effect (Study 1), and systematically explore its underlying mechanisms (Study 2).
In Study 1, we recruited 64 U.S. and 75 Chinese participants, hypothesizing that the “person-based” mere ownership effect would be more salient among Easterners than Westerners, as Easterners often require an external motivation to self-enhance (Yamagishi et al.,2012). All participants were manipulated with a self-threat related to their culturally significant ability. For instance, informing Chinese (US) participants that they lacked holistic thinking (assertiveness). In each culture, participants received either holistic oil, assertive oil, or no oil (control). Results showed that U.S. participants reported higher self-enhancement than Chinese participants. But when Chinese participants possessed the holistic oil (external motivation), they reported a comparable level of self-enhancement as their US counterparts.
Study 2 aimed to replicate and extend the findings of Study 1 by embedding cultural mindset priming within a cultural identification context. Based on prior research showing that Hong Kong participants can shift between Eastern and Western mindsets, we recruited 164 bilingual participants and randomly assigned them to either an Eastern or Western cultural mindset priming condition. As expected, participants primed with an Eastern mindset reported higher holistic efficacy than those primed with a Western mindset, even without relying on external objects. This suggests that salient cultural identification provided a culturally congruent justification for self-enhancement. However, results from the Western-primed condition were more complex. Unlike the U.S. participants in Study 1, Hong Kong participants primed with a Western mindset did not show consistently higher self-enhancement. Instead, participants primed with a Western mindset showed lower holistic efficacy than those primed with an Eastern mindset, and both mindset groups had similarly low assertive efficacy. Given the ongoing social and cultural changes in Hong Kong, previously effective priming tasks may no longer yield the same results, therefore, these findings suggest that Western mindset priming may have led to identification confusion, making it difficult for participants to adopt a clear self-enhancement strategy
The impact of housing stress on fertility intentions among young people in urban China
No full textHousing has been recognised as an important factor associated with decreasing fertility in urban China. Despite a growing number of studies dedicated to the investigation of the impact of adverse housing conditions on fertility intentions, the understanding of housing-fertility nexus remains relatively fragmented, characterised by an incomplete picture of housing stressors and conflicting findings that fail to explain why some people experiencing adverse housing conditions tend to have lower fertility intentions whilst others do not. Framing the transactional theory of stress into the housing contexts, this study redefined housing stress to enhance its explanation compared with correlation explanations in the previous literature. Using Shenzhen as a study area, this study aimed to address the research gaps by (1) exploring the multidimensional housing stressors that shape young people’s fertility intentions and (2) examining the mechanism of how the stressor is perceived as stress, and how the stress influences fertility intentions.
Employing a design combining grounded theory and fuzzy-set qualitative comparative analysis, this study identified market-driven, institutional, and household-driven stressors, revealing their dynamics interplay. It elucidated dual impact pathways of housing stress on fertility intentions. Specifically, the transactional interaction between housing stressors and ideal-oriented beliefs fosters a threat appraisal, suppressing fertility intentions, while those with reality-oriented beliefs promotes a challenge appraisal, resulting in a sustaining effect on fertility intentions. The sufficiency analysis of fsQCA identified four distinct configuration types, highlighting causal complexity around these pathways. By exploring the nexus between housing stress and fertility intentions, this study provides insights to mitigate housing stress and fertility anxiety among young people in urban China
《火種》中的上海敘事與工人階級的情感書寫
No full text艾明之的《火種》為「十七年」文學（1949-1966年中國文學時期）的重要文學作品之一，其記載了上海工人於一九一〇至一九二〇年代末，在中國的一系列政治形勢下的鬥爭生活。作者更以親歷者的身分描繪了半殖民地時期的上海，並細膩刻畫蘇州河畔工人的生活概況，呈現底層人民的人性光輝。
本文將進行文本細讀，並與史料及同類型文學作品相互參照。全文共分為五個章節，第一章為緒論，說明研究動機和《火種》的學術價值，以及梳理相關研究成果並予以評析。第二章選取文本中的上海敘事，細看作者如何呈現一九一〇至一九二〇年代末的社會風貌。第三章則從兩方面檢視工人的工作待遇，分別是工人在上海租界以外的生活，以及工人離開上海後，在煤礦的處境，從而全面分析工人階級的生活狀態。第四章則透過敘述柳金松和殷玉花、以及小廣東黃阿康和金鳳的愛情，探討他們純樸的品格與豐富的情感世界。最後，在第五章進行總結，再次闡述《火種》在當代文學史中的多重意義
甲骨文“本”字試解
No full text甲骨文的“本”字，目前學界有釋“垂”、釋“叕”、釋“瓜”等觀點，這些觀點在字形、辭例釋讀方面都存在一些缺陷。本文從“本”的字形特徵入手進行分析，認爲該字應當是“本”之初文，在相關祭祀卜辭中應理解爲人牲的軀榦或下部，或與商代的肢解或腰斬之刑有關。
At present, there are various viewpoints about the “本” character Oracle Bone Inscription, such as the interpretation of “chui 垂”, “zhui 叕”, and “gua 瓜”, these viewpoints have some shortcomings in terms of form and example interpretation. This paper analyzes the form characteristics of “本” and believes that this character should be the initial form of “ben 本”. In related oracle inscriptions, it should be understood as the torso or lower part of a human sacrifice, it may be related to the dismemberment or beheading punishment in the Shang Dynasty
“歌詩必類”的禮制語境與禮義、樂義
No full text“歌詩必類”的本義，當求之於歷史語境。晉平公在繼位之初與諸侯會盟時，欲宣示晉爲盟主，參與的諸侯大夫皆以朝聘之禮見之。其中的“歌詩必類”，非諸侯引詩、賦詩的探求詩義，而是以大夫見天子之禮歌之。齊欲與晉爭霸，齊使高厚歌不類，引發晉伐齊。故“歌詩必類”的禮義，是依頌、雅、風的規格“上取下就”而用之，用所歌篇章的樂義交流，此爲春秋諸侯大夫歌詩的禮樂規則。其意義在於歌唱的詩篇要取其禮義、樂義，而非詩義。
The original meaning of “The Odes chanted must fit proper categories” should be sought in the historical context. When Duke Ping of Jin first held a meeting with the princes after he ascended the throne, he wanted to declare Jin as the leader of the alliance, and all the princes and ministers who participated in the meeting met him with the ceremony of paying homage. “The Odes chanted must fit proper categories” here does not mean that the princes quote poems or compose poems to explore the meaning of the poems, but that they are sung according to the etiquette of officials meeting the emperor. Qi wanted to contend for hegemony with Jin, so Qi’s envoy Gao Hou sang an unconventional song, which led to Jin’s attack on Qi. Therefore, the etiquette and morality of “The Odes chanted must fit proper categories” is to “take from the top and follow the bottom” according to the standards of Song, Ya and Feng, and use the music meaning of the chapters being sung to communicate. This is the etiquette and music rules for singing poems by princes and officials in the Spring and Autumn Period. Its significance lies in that the poems sung should be based on their ritual and musical meaning rather than their poetic meaning
漢文帝朝禮制建設與君臣政治博弈
No full text漢文帝登位後，與高祖時形成的權臣勢力，尤其是擁立他的大臣，無可避免地形成錯綜複雜的矛盾。文帝試圖通過各種方式消解權臣的壓力，其中一個重要方面，是通過祭祀制度、漢朝德運、立五帝廟等禮制上的改革和建設，强化其個人繼位的合法性，並藉此鞏固和擴張個人的政治權力。權臣則以不同形式、在不同程度上，對文帝進行多方制約甚至反抗，其根本目的在於阻止文帝的權力擴張，保護權臣既有的政治權力和利益。聚觀多次的交鋒，基本以文帝落敗，權臣慘勝告終。而文帝的政治合法性問題，最終由劉啓以太子身份順利繼位爲景帝，恢復漢朝帝皇繼位的制度化模式纔得以解決。
After ascending to the throne, Emperor Wen faced complex conflicts with the powerful courtiers who had supported Han Gaozu, especially those who were his supporters. He attempted to alleviate the pressure from these ministers through various means and reinforce the legitimacy of his succession. Emperor Wen aimed to consolidate and expand his personal political power through ritual reforms and constructions, such as the ritual system, the Han Dynasty’s virtue, and the establishment of the Five Emperors’ Temples. However, powerful courtiers restrained and rebelled against the emperor in various ways to prevent the expansion of his power and protect their own vested interests. The confrontations resulted in the defeat of Emperor Wen but the courtiers emerged victorious at a high cost. The political legitimacy of Emperor Wen was not resolved until Liu Qi was successfully enthroned as the Crown Prince with the institutionalized royal succession in the Han Dynasty
The changing dynamics of India-China relations since 2014 : a shift from asymme tric to symmetric security dilemma?
No full textThe new global geopolitical order has shifted from the Atlantic to the Indo-Pacific at the dawn of the twenty-first century. Given the fact that both China and India are currently regarded as major rising powers, the relationship between the two countries is among the most important concerns in the contemporary geopolitical landscape. Therefore, it is of utmost significance to have a comprehensive understanding of the dynamics of their relationship. The dynamics of the relationship between China and India have been marked by the presence of a security issue for a considerable amount of time. It has been suggested that the security dilemma has been a cornerstone of their overall relation. However, the dynamics of the security dilemma between both the countries has been asymmetrical. This is mainly due to the fact that there was asymmetry in power and how threat was perceived.
Beginning in the 1980’s, as the power relations between the two states became more unequal due to China\u27s rapid economic ascent, it rendered the threat perception decidedly one sided, thus creating an asymmetrical security dilemma. However, this research finds that with India\u27s economic rise over the past two decades has seen a sharp shift in its policy approach towards China, particularly since the emergence of Narandra Modi as India’s prime minister in 2014. Specifically, there has been a clear shift in the dynamics of the security dilemma and the general threat perception. Since 2014, India\u27s approach to China has been far more robust making it less asymmetrical, as evidenced by the border crisis in 2017 in Doklam and 2020 at Galwan. Similarly, New Delhi’s Indo-Pacific strategy apparently designed to balance China have also contributed to the improvement of its relations with third parties, most notably the United States. Additionally, India has been sharpening its peripheral strategy underscoring its confidence vis-à-vis direct security competition with China. Overall, this has made the perception of the threat, which was previously emanating from India, to become more symmetrical. This, in turn, is also causing the security dilemma, which was previously unduly asymmetric in nature, to become more symmetric
大數據對物質文化史的衝擊 : 以古建築及賦税用語“間”爲例
No full text中國古代建築常以“間”或“半間”計數，然學界少見深入研究。本文即利用目前已建置完成之百餘億字可全文檢索的大數據新環境，查找出自唐宋以迄明清留有長闊或高度的近8,000間記事，並嘗試透過這些事例，深化對此古建築用語的相關認識。發現“間”字用在地舍的具體描述或租税計算時，其定義並不相同。前者對個别房舍通常只計以“間”或“半間”，取建築物正面“兩柱夾覆之中”爲1間，並以間闊1丈爲“標準值”，但往往略有增減。後者則是將面闊1間之房舍依其進深、屋況或地段等因素綜合估算（金代乃以“椓”數表之），再换算成間數（遂有畸零之數）以徵收租税。此一新認知展現大數據對傳統物質文化史研究所具有的啓發意義。
Ancient Chinese architecture is often counted by “Jian” or “Half Jian”, but there is little in-depth research in academia. This article uses the new big data environment that has been built with more than 10 billion words of full-text search to search for nearly 8, 000 records of length, width or height from the Tang and Song Dynasties to the Ming and Qing Dynasties, and attempts to deepen the understanding of this ancient architectural terminology through these examples. It was found that the definition of the word “Jian” was different when it was used in the specific description of a place or in the calculation of rent and taxes. The former usually only counts individual houses as “rooms” or “half rooms”, with the area between two pillars on the front of the building being considered one room, and the width of the room being 10 feet as the “standard value”, but there are often slight increases or decreases. The latter is to make a comprehensive estimate of the depth, condition of the house or location of a house with a width of one room (in the Jin Dynasty, it was expressed as the number “Zhuo”), and then convert it into the number of rooms (thus there are odd numbers) to collect rent and taxes. This new understanding demonstrates the inspiring significance of big data for the study of traditional material culture history