Search

Multi-well extraction is a prevalent technique in coalbed methane (CBM) recovery projects. Although numerous studies have extensively explored aspects such as well spacing, the degree of multi-well pumping, and well count, the dynamics of fracture microstructure evolution in proximity to wells—particularly in inter-well regions—remain inadequately understood in relation to the effects of multi-well mining project. This research delved into the multi-well extraction methodology employed in CBM recovery endeavors, aiming to elucidate the development of the fracture microstructure network. We introduce a novel, interdisciplinary, and integrative research framework that amalgamates the multi-field coupling effects observed during the multi-well extraction process with fractal theory. This model has been validated, and it facilitates the examination of changes in fracture micro-evolution subjected to multi-well extraction. Additionally, this study investigated alterations in fracture characteristics, seam stress, and CBM pressure within sensitive zones (i.e., inter-well spaces and adjacent areas) under varying extraction pressures. Following a 180-day extraction period, the findings indicate a significant reduction in gas pressure by 83.9% for the extraction wells and the nearby areas, alongside a decrease in fracture network length by 10.94% and density by 5.04%. Compared to existing models for assessing multi-well CBM extraction, our interdisciplinary model demonstrates considerable analytical superiority. Notably, when the fractal parameters Df and DTf, which characterize fracture density and tortuosity quantitatively, increase from 1.2 to 1.8, the residual gas pressure is reduced further by 11.6% and increased further by 3.9%, respectively. [ABSTRACT FROM AUTHOR]

The underlying atomistic mechanism of deformation is a central problem in mechanics and materials science. Whereas deformation of crystalline metals is fundamentally understood, the understanding of deformation of amorphous metals lacks behind, particularly identifying the involved temporal and spatial scales. Here, we reveal that at small scales the size-dependent deformation behavior of amorphous metals significantly deviates from homogeneous flow, exhibiting increasing deformation rate with reducing size and gradually shifted composition. This transition suggests the deformation mechanism changes from collective atomic transport by viscous flow to individual atomic transport through interface diffusion. The critical length scale of the transition is temperature dependent, exhibiting a maximum at the glass transition. While viscous flow does not discriminate among alloy constituents, diffusion does and the constituent element with higher diffusivity deforms faster. Our findings yield insights into nano-mechanics and glass physics and may suggest alternative processing methods to epitaxially grow metallic glasses. How matter deform is a central question in mechanics of materials science. Here, the authors reveal a size-dependent deformation in amorphous metals changing at ~100 nanometer sample size from collective homogeneous flow to interface diffusion realized through individual atomic transport. [ABSTRACT FROM AUTHOR]

Key message: This study identified a novel SNP and developed a highly efficient KASP marker for drought tolerance in wheat by genotyping NILs targeting a major QTL for drought tolerance using an SNP array and validation with commercial varieties. Common wheat (Triticum aestivum L.) is an important winter crop worldwide and a typical allopolyploid with a large and complex genome. With global warming, the environmental volatility and incidence of drought in wheat-producing areas will increase. Molecular markers for drought tolerance are urgently needed to enhance drought tolerance breeding. Here, we genotyped four near-isogenic line (NIL) pairs targeting a major QTL qDSI.4B.1 on wheat chromosome arm 4BS for drought tolerance using the 90K SNP Illumina iSelect array and discovered a single nucleotide polymorphism (SNP) (Excalibur_c100336_106) with consistent genotype–phenotype associations among all four NIL pairs and their parents. Then, we converted the SNP into a Kompetitive Allele-Specific PCR (KASP) marker, with an accuracy of 100% for the four NIL pairs and their parents and as high as 81.8% for the 44 tested wheat lines with known phenotypes collected from Australia and China. Two genes near this SNP were suggested as candidate genes for drought tolerance in wheat after checking the Chinese Spring reference genome annotation version 1.1. One gene, TraesCS4B02G085300, encodes an F-box protein reportedly related to the ABA network, a main pathway for drought tolerance, and another gene, TraesCS4B02G085400, encodes a calcineurin-like metallophos-phoesterase transmembrane protein, which participates in Ca2+-dependent phosphorylation regulatory system. Based on this work and previous research on pre-harvest sprouting, we established a quick and efficient general SQV-based approach for KASP marker development, integrating genotyping by SNP arrays (S) using NILs targeting major QTL for a specific trait (Q) and validating them with commercial varieties (V). The identified SNP and developed KASP marker could be applied to marker-assisted selection in drought breeding, and further study of the candidate genes may improve our understanding of drought tolerance in wheat. [ABSTRACT FROM AUTHOR]

The prevalence of online user-generated content has attracted great interest in textual sentiment analysis, which provides a low-cost yet effective way to discern consumers and markets. A mainstream of sentiment analysis is to construct a classification model with Bag-of-Words (BoW) features, but the large vocabulary base and skewed distribution of term frequency consistently pose research challenges, which is made even worse by the limited valid sentiment labels. In light of this, in this paper, we propose a novel method called Structural Holes based Sentiment Classifier (SHSC) for BoW-based sentiment classification. The key to SHSC is to reinforce the classification contribution of semantically rich words with clear-cut sentiment polarity. To this end, a word co-occurrence network is carefully constructed to represent both high and low frequency words. The work to find classification-inefficient words is then transformed into the identification of so-called bridge nodes that occupy the positions of structural holes in the network. Two interesting measures, i.e., information advantage rank and control advantage weight, are then designed elaborately for this purpose, which are based on the proposed sentiment-label propagation and short-path computation algorithms, respectively. SHSC finally feeds this information as the key regularizers into a simple regression model to guide parametric learning. Extensive experiments on real-world text datasets demonstrate the advantage of our SHSC model over competitive benchmarks, particularly when sentiment labels are scarce. The effectiveness of uncovering structural holes for sentiment classification is also carefully verified with some robustness checks and demonstration cases. [ABSTRACT FROM AUTHOR]

Detecting anomalies from a massive amount of user behavioral data is often liken to finding a needle in a haystack. While tremendous efforts have been devoted to anomaly detection from temporal graphs, existing studies rarely consider community evolution and evolutionary paths simultaneously, and analyze those characteristics for the purpose of anomaly detection. Therefore, we propose a two-stage anomaly detection (TSAD) framework to detect anomalies. In this study, we suggest detecting the community evolution events from a sequence of snapshot graphs by constructing an evolution bipartite graph and designing community similarity scores. We then propose a novel anomaly detection method combining community evolution-based anomaly detection and evolutionary path-based anomaly detection. An anomalous score is designed to detect anomalous community evolution events by extracting the characteristics of evolution communities in the community evolution-based anomaly detection method. Moreover, to reduce the false alarm rate, we propose evolutionary path-based anomaly detection to further detect the abnormality of the identified normal evolutionary paths by extracting the characteristics of the identified anomalous evolutionary paths based on community evolution-based anomaly detection. We conduct extensive experiments on real-world datasets and demonstrate that TSAD consistently outperforms competitive baseline methods in anomaly detection. [ABSTRACT FROM AUTHOR]

During gas drainage process, addressing coal and gas outbursts—a major mining hazard—requires understanding the multi-field coupling effects on rock solid damage and the overlooked role of pore structure evolution. We have developed an advanced model integrating these dynamics, enhancing traditional fractal theory for porous media to incorporate multi-field effects crucial in gas drainage. This model, validated through finite element analysis software and field data, offers a microscopic view of the extraction process, analysing stress, deformation, and damage evolution within the coal seam. Notably, increasing pore numbers significantly affect gas pressure (up to 19.7% variation) and coal seam displacement (increasing by 50.9%), with gas production rising by up to 17.6%. Additionally, coupled multi-field and extraction conditions result in a 9.8% increase in maximum pore diameter. Our findings highlight that the evolution of pore properties in coal seams is a more critical factor in outburst risk than effective stress or coal deformation. This novel, interdisciplinary fractal model is versatile, applicable to various gas drainage scenarios, and offers insights for optimising production under diverse coal conditions. [ABSTRACT FROM AUTHOR]

The types of land used for farmland can greatly influence the source and accumulation risk of heavy metals in soil. However, the apportioning quantitatively the source of soil heavy metals has been studied insufficiently, especially in terms of different types of farmland. In this study, a total of 252 soil samples were taken from dry land, paddy fields and greenhouse fields in the Jinyuan district of Taiyuan city, China, to assess the accumulation risk of heavy metals (As, Cd, Cr, Cu, Hg, Ni, Pb and Zn). The results were then integrated, and source apportionment was evaluated by geospatial analysis, multivariate statistical analysis and positive matrix factorization (PMF). Cr, Cd and Hg were the dominant pollutants in the studied area. Accumulation risk by Cd and Cu was more severe in greenhouse fields than in dry land or paddy fields, whereas As, Hg and Pb had relatively higher accumulation in paddy fields than in dry land or greenhouse fields. Hg was derived mainly from coal combustion by atmospheric precipitation for the three types of farmland. Long-term irrigation using sewage is the main reason for the accumulation of Cu and Ni in dry land soil, Cu and Zn in paddy field soil and Zn in greenhouse soil. Cd in dry land, Cd and Pb in paddy fields and Cd, Cu, Ni and Pb in greenhouse fields were primarily added to soil through fertilization. Sewage irrigation and fertilization were the dominant sources of heavy metals for paddy field (31.3%) and greenhouse field (33.1%), respectively. [ABSTRACT FROM AUTHOR]

Traditional techniqueset identification, we developed GraphDTI, a robust machine learning framework integrating the molecular-level information on drugs, proteins, and binding sites with the system-level information on gene expression and protein-protein interactions. In order to properly evaluate the performance of GraphDTI, we compiled a high-quality benchmarking dataset and devised a new cluster-based cross-validation p to identify macromolecular targets for drugs utilize solely the information on a query drug and a putative target. Nonetheless, the mechanisms of action of many drugs depend not only on their binding affinity toward a single protein, but also on the signal transduction through cascades of molecular interactions leading to certain phenotypes. Although using protein-protein interaction networks and drug-perturbed gene expression profiles can facilitate system-level investigations of drug-target interactions, utilizing such large and heterogeneous data poses notable challenges. To improve the state-of-the-art in drug targrotocol. Encouragingly, GraphDTI not only yields an AUC of 0.996 against the validation dataset, but it also generalizes well to unseen data with an AUC of 0.939, significantly outperforming other predictors. Finally, selected examples of identified drug-target interactions are validated against the biomedical literature. Numerous applications of GraphDTI include the investigation of drug polypharmacological effects, side effects through off-target binding, and repositioning opportunities. [ABSTRACT FROM AUTHOR]

Direct measurement of critical cooling rates has been challenging and only determined for a minute fraction of the reported metallic glass forming alloys. Here, we report a method that directly measures critical cooling rate of thin film metallic glass forming alloys in a combinatorial fashion. Based on a universal heating architecture using indirect laser heating and a microstructure analysis this method offers itself as a rapid screening technique to quantify glass forming ability. We use this method to identify glass forming alloys and study the composition effect on the critical cooling rate in the Al–Ni–Ge system where we identified Al51Ge35Ni14 as the best glass forming composition with a critical cooling rate of 104 K/s. [ABSTRACT FROM AUTHOR]

In recent years, short text topic modeling has drawn considerable attentions from interdisciplinary researchers. Various customized topic models have been proposed to tackle the semantic sparseness nature of short texts. Most (if not all) of them follow the bag-of-words assumption, which, however, is not adequate since word order and phrases are often critical to capturing the meaning of texts. On the other hand, while some existing topic models are sensitive to word order, they do not perform well on short texts due to the severe data sparseness. To address these issues, we propose the Pseudo-document-based Topical N-Grams model (PTNG), which alleviates the data sparsity problem of short texts while is sensitive to word order. Extensive experiments on three real-world data sets with state-of-the-art baselines demonstrate the high quality of topics learned by PTNG according to UCI coherence scores and more discriminative semantic representation of short texts according to classification results. [ABSTRACT FROM AUTHOR]

Fraud detection is usually compared to finding a needle in a haystack and remains a challenging task because fraudulent acts are buried in massive amounts of normal behavior and true intentions may be disguised in a single snapshot. Indeed, fraudulent incidents usually take place in consecutive time steps to gain illegal benefits, which provides unique clues for probing fraudulent behavior by considering a complete behavioral sequence rather than detecting fraud from a snapshot of behavior. Additionally, fraudulent behavior may involve different parties, such that the interaction patterns between sources and targets can help distinguish fraudulent acts from normal behavior. Therefore, in this paper, we model the attributed behavioral sequences generated from consecutive behaviors in order to capture the sequential patterns, while those that deviate from the pattern can be detected as fraudulence. Considering the characteristics of the behavioral sequence, we propose a novel model, NHA-LSTM, by augmenting the traditional LSTM with a modified forget gate, where the interval time between consecutive time steps is considered. Furthermore, we design a self-historical attention mechanism to allow for long time dependencies, which can help identify repeated or cyclical appearances. In addition, we propose an enhanced network embedding method, FraudWalk, to construct embeddings for the nodes in the interaction network with regard to higher-order interactions and particular time constraints for revealing potential group fraudulence. The node embeddings, along with the feature vectors, are fed into the model to capture the interactions between sources and targets. To validate the effectiveness of sequential behavior embeddings, we experiment on a real-world telecommunication dataset with prediction and classification tasks based on the learned embeddings. The experimental results show that the learned embeddings can better identify fraudulent behavior. Finally, we visualize the weights of the attention mechanism to provide a rational interpretation of human behavioral patterns. [ABSTRACT FROM AUTHOR]

The occurrence of plastic particles in aquatic environment has led to enormous concern in the past few years. The sorption behaviors of harmful organic compounds by plastic particles can increase their concentrations by several orders of magnitude influencing their global transport in the marine environment. Five types of mesoplastics (5–20 mm) and five types of microplastics (< 5 mm) were selected to investigate the sorption behaviors of three typical organic compounds (phenanthrene, nitrobenzene, and naphthalene). For phenanthrene, most microplastics have stronger sorption ability than that of mesoplastics due to the higher specific surface area (SSA). However, the sorption ability of nitrobenzene on low-density polyethylene (LDPE) mesoplastics was higher than that on LDPE microplastics, and the sorption ability of naphthalene on polyvinyl chloride (PVC) mesoplastics was higher than that on PVC microplastics, which were attributed to the presence of functional groups on the surface of mesoplastics, induced by adding slip agents, lubricant, plasticizer, stabilizer, etc. during film production. Talcum-filled polypropylene (PP) microplastics had strongest sorption ability to nitrobenzene and naphthalene due to the presence of talcum and high SSA. For unmodified microplastics, the sorption abilities of phenanthrene, nitrobenzene, and naphthalene were all followed the order of high-density polyethylene (HDPE) > polystyrene (PS) > LDPE > PVC after SSA normalization. Thus, SSA and the functional groups on the surface of plastic particles should be considered when the sorption behaviors of harmful organic compounds on plastic particles are studied. [ABSTRACT FROM AUTHOR]

The physicochemical properties and heavy metal(loid) concentrations of the river water both fluctuate greatly along the river affected by mining activities, and the transportation of heavy metal(loid)s is therefore more complicated than unpolluted river. Dissolved and particulate heavy metal(loid)s in a river polluted by mining activities were measured to study their temporal-spatial variation and partitioning. The concentrations of dissolved arsenic (As), cadmium (Cd), manganese (Mn), nickel (Ni), lead (Pb), and zinc (Zn) were considerably high at the sites near the mine area. Notably, dissolved As at most sites were higher than the Chinese quality criterion of class II for surface water indicating high environmental risk. Mn and Pb at most sites and Ni at a part of the sites mainly existed in the particulate phase. For other heavy metal(loid)s, i.e., As, Cd, chromium (Cr), and Zn, the particulate phase was extremely high at the sites near the mine area and responsible for heavy metal(loid) transport. Significant correlations between particulate heavy metal(loid)s and temperature and electrical conductivity (EC) were found. However, the partitioning of heavy metal(loid)s did not significantly relate to the river water properties, due to most heavy metal(loid)s in suspended particulate matter (SPM) are stable and affected less by water properties. Except for Cr and Ni, other heavy metal(loid)s showed high concentrations in sediments, and considerable Cd, Mn, and Zn existed in exchangeable and carbonate fraction indicating high environmental risk. The environmental assessment of SPM showed that Cd, Zn, and As, as the main pollutants in SPM, all reached extremely polluted level at the sites near the mine area, and the environmental risk of heavy metal(loid)s in SPM was higher during dry season than that during wet season. The results can contribute to understanding the partitioning and transportation of heavy metal(loid)s in the river affected by mining activities. [ABSTRACT FROM AUTHOR]

In the present study, Fe nanoparticles (NPs) can ignite when exposed to a conventional camera flash. The ignition process of Fe NPs was composed of the initial ignition stage with the maximum temperature level of 2000 K and the burning stage with temperature level of 800 K. The microstructure characterization indicated Fe NPs were oxidized to FeO via the exposure-melt mechanism. It was found that more particle numbers per unit area can lead to lower minimum ignition energy which may be caused by the enhancement light energy absorption. The light energy absorbed from the flash was influenced by wavelengths but the conversion of Fe was only related to the packing mass. [ABSTRACT FROM AUTHOR]

Purpose: The size of soil particles strongly affects the accumulation and adsorption of heavy metals which partly controls the co-transport of heavy metals by soil colloids. However, the effect of the size of soil particles on the accumulation and adsorption of heavy metals in the colloidal dimension has seldom been studied. In this study, variable charge soils were selected and separated into five size fractions to elucidate the effect of the size of soil particles on Cd accumulation and adsorption. Materials and methods: Five soil particle size fractions (>10, 10-1, 1-0.45, 0.45-0.2 and <0.2 μm) were obtained from Cd-contaminated soil by natural sedimentation and fractional centrifugation. The concentrations and species of Cd were measured in various sized soil particles. Batch adsorption experiments of Cd on the obtained soil particles were conducted under different pH values and concentrations of NaCl. Results and discussion: Generally, the concentration of Cd increased with decreasing soil particle sizes, and the Cd proportion of exchangeable and carbonate fraction decreased from 43.84 to 17.75% with decreasing particle size. The soil particles with a size of 10-1 and <0.2 μm possessed a stronger adsorption ability than the other fractions in most cases. Moreover, the Cd adsorption capacities of the soil particles increased with increasing pH values and decreasing concentrations of NaCl, especially for soil particles containing more organic matter (OM) and variable charge minerals. Conclusions: Smaller soil particles are more capable of accumulating Cd and make Cd more stable. The adsorption capability of Cd is negatively related to the particle size and NaCl concentration and is positively related to the pH. The effects of the size of variable charge soil particles on Cd accumulation and adsorption are attributed to the differences in the physicochemical properties among various soil particle size fractions. This study contributes to the understanding of the co-transport of heavy metals in soil by soil colloids. [ABSTRACT FROM AUTHOR]

Combustion characteristics of nanofluid fuels containing aluminum nanoparticles were investigated in half-opening slot tubes from the fundamental view. The effects of particle loading rates (0.25% and 2.5% by weight), type of base fuels (ethanol and butanol), and fuel flow rates (0.2, 0.6, and 1 mL/min) were studied in details. The combustion characteristics of the nanofluid fuels and pure based fuels were also examined to provide a comparison. Flame was unstable with reignition, stable state, nearly extinguishment repeatedly at low flow rate. At medium flow rate, flame height was increased and flame tended to be stable. At high flow rate, flame became unstable and was disturbed by the droplet forming and dripping significantly. Al atoms inside the oxide layer should be melted before the particles combustion, while Al oxide layer should be melted before the particles aggregates combustion. The effects of particles on the combustion characteristics, especially on the evaporation rate of base fuel, were discussed. The reasons for various combustion phenomena of nanofluid fuels were given, which can provide the useful guidance for the experimental research and practical applications of nanofluid fuels. [ABSTRACT FROM AUTHOR]

Heavy metal contamination of soils has been a long-standing environmental problem in many parts of the world, and poses enormous threats to ecosystem and human health. Speciation of heavy metals in soils is crucial to assessing environmental risks from contaminated soils. In this study, total concentrations and speciation of As, Cd, Cr, Cu, Mn, Ni, Pb, and Zn were measured for agricultural soils near mines along the Diaojiang River in Guangxi Zhuang Autonomy Region, China. The sources of heavy metals in soils also were identified to assess their effect on speciation distribution of soil heavy metals. Furthermore, the speciation distribution of Cd and Zn, main soil heavy metal pollutants, in dry land and paddy soils were compared. Results showed that there were two severely polluted regions near mine area reaching alarming pollution level. As, Cd, Pb, and Zn were more affected by mining activities, showing very strong pollution level in soils. The mean percentage of exchangeable and carbonate fraction was highest and up to 46.8 % for Cd, indicating a high environmental risk. Greater bioavailable fractions of As, Cd, Cu, Mn, Pb, and Zn were found in soils heavily polluted by mining activities, whereas Cr and Ni as geogenic elements in the stable residual fraction. In addition, in the dry land soils, reducible fraction proportion of Cd was higher than that in the paddy soils, whereas exchangeable and carbonate fraction of Cd was lower than that in the paddy soils. Oxidizable fraction of Zn was higher in the paddy soils than that in the dry land soils. The results indicate that the sources of soil heavy metals and land types affect heavy metal speciation in the soil and are significant for environmental risk assessment of soil heavy metal pollutions. [ABSTRACT FROM AUTHOR]

Organic phosphorus (OP) species in sediments of the Baiyangdian Lake in China was investigated via fractionation and phosphorus-31 nuclear magnetic resonance (P NMR) spectroscopy. Results of chemical fractionation showed that different OP fractions ranked in the order: humic acid-P > HCl-OP > fulvic acid-P ≈ NaHCO-OP > residual OP. Labile OP and moderately labile OP, which can be degraded for phytoplankton uptake, contributed to 58.7-68.5% of total extracted OP, indicative of the potential P release from sediments in the lake. P NMR spectroscopy results suggested the rank order of P species present in the NaOH-EDTA extractant of the sediments: orthophosphate > monoester-P > DNA-P > pyrophosphate-P. Phytate, polyphosphates, and phosphonates, which appeared in sediments of some deeper lakes, were not detected in the shallow lake. Significant correlations were identified between total P (TP) in water column and sediment TP, monoester-P and DNA-P, positively indicating that sediment P species, especially OP components, should probably contribute to the contents of P in water column and further to the lake eutrophication. [ABSTRACT FROM AUTHOR]

Jujube witches' broom (JWB) is a destructive disease, caused by a phytoplasma, of Chinese jujube ( Ziziphus jujuba Mill.), an important traditional fruit in China and South Korea. Micrografting of plantlets to detect those affected with JWB is difficult due to differences in stem diameter between the scion and rootstock. In this study, an in vitro micrografting technique, reverse-cleft graft, suitable for rootstock thinner than scion was developed and optimized. The JWB phytoplasma was successfully transmitted into the scion from diseased rootstock using this technique as determined by polymerase chain reaction. Micrografted plantlets of Chinese jujube were incubated on Murashige and Skoog medium without plant growth regulators. The survival frequencies of 78-91%, depending on grafting combinations, were obtained. [ABSTRACT FROM AUTHOR]

Next location prediction has aroused great interests in the era of internet of things (IoT). With the ubiquitous deployment of sensor devices, e.g., GPS and Wi-Fi, IoT environment offers new opportunities for proactively analyzing human mobility patterns and predicting user's future visit in low cost, no matter outdoor and indoor. In this paper, we consider the problem of next location prediction in IoT environment via a session-based manner. We suggest that user's future intention in each session can be better inferred for more accurate prediction if patterns hidden inside both trajectory and signal strength sequences collected from IoT devices can be jointly modeled, which however existing state-of-the-art methods have rarely addressed. To this end, we propose a trajectory and sIgnal sequence (TSIS) model, where the trajectory transition regularities and signal temporal dynamics are jointly embedded in a neural network based model. Specifically, we employ gated recurrent unit (GRU) for capturing the temporal dynamics in the multivariate signal strength sequence. Moreover, we adapt gated graph neural networks (gated GNNs) on location transition graphs to explicitly model the transition patterns of trajectories. Finally, both the low-dimensional representations learned from trajectory and signal sequence are jointly optimized to construct a session embedding, which is further employed to predict the next location. Extensive experiments on two real-world Wi-Fi based mobility datasets demonstrate that TSIS is effective and robust for next location prediction compared with other competitive baselines. [ABSTRACT FROM AUTHOR]

The prevalence of online media has attracted researchers from various domains to explore human behavior and make interesting predictions. In this research, we leverage heterogeneous data collected from various online platforms to predict Taiwan’s 2016 general election. In contrast to most existing research, we take a “signal” view of heterogeneous information and adopt the Kalman filter to fuse multiple signals into daily vote predictions for the candidates. We also consider events that influenced the election in a quantitative manner based on the so-called event study model that originated in the field of financial research. We obtained the following interesting findings. First, public opinions in online media dominate traditional polls in Taiwan election prediction in terms of both predictive power and timeliness. But offline polls can still function on alleviating the sample bias of online opinions. Second, although online signals converge as election day approaches, the simple Facebook “Like” is consistently the strongest indicator of the election result. Third, most influential events have a strong connection to cross-strait relations, and the Chou Tzu-yu flag incident followed by the apology video one day before the election increased the vote share of Tsai Ing-Wen by 3.66%. This research justifies the predictive power of online media in politics and the advantages of information fusion. The combined use of the Kalman filter and the event study method contributes to the data-driven political analytics paradigm for both prediction and attribution purposes. [ABSTRACT FROM AUTHOR]