User-generated reviews on the Web reflect users’ sentiment about products, services and social events. Existing researches mostly focus on the sentiment classification of the product and service ...reviews in document level. Reviews of social events such as economic and political activities, which are called social reviews, have specific characteristics different to the reviews of products and services. In this paper, we propose an unsupervised approach to automatically discover the aspects discussed in Chinese social reviews and also the sentiments expressed in different aspects. The approach is called Multi-aspect Sentiment Analysis for Chinese Online Social Reviews (MSA-COSRs). We first apply the Latent Dirichlet Allocation (LDA) model to discover multi-aspect global topics of social reviews, and then extract the local topic and associated sentiment based on a sliding window context over the review text. The aspect of the local topic is identified by a trained LDA model, and the polarity of the associated sentiment is classified by HowNet lexicon. The experiment results show that MSA-COSR cannot only obtain good topic partitioning results, but also help to improve sentiment analysis accuracy. It helps to simultaneously discover multi-aspect fine-grained topics and associated sentiment.
Long short-term memory networks (LSTMs) have gained good performance in sentiment analysis tasks. The general method is to use LSTMs to combine word embeddings for text representation. However, word ...embeddings carry more semantic information rather than sentiment information. Only using word embeddings to represent words is inaccurate in sentiment analysis tasks. To solve the problem, we propose a lexicon-enhanced LSTM model. The model first uses sentiment lexicon as an extra information pre-training a word sentiment classifier and then get the sentiment embeddings of words including the words not in the lexicon. Combining the sentiment embedding and its word embedding can make word representation more accurate. Furthermore, we define a new method to find the attention vector in general sentiment analysis without a target that can improve the LSTM ability in capturing global sentiment information. The results of experiments on English and Chinese datasets show that our models have comparative or better results than the existing models.
Stock price prediction is an important and complex time-series problem in academia and financial industries. Stock market prices are voted by all kinds of investors and are influenced by various ...factors. According to the literature studies, such as Elliott’s wave theory and Howard’s market cycle investment theory, the cyclic patterns are significant characteristics of the stock market. However, even several studies that do consider cyclic patterns (or similar concepts) suffered from the data leakage or boundary problems, which could be impractical for real applications. Inspired by the abovementioned, we propose a hybrid deep learning model called mWDN-LSTM, which correctly utilizes the cyclic patterns’ information to predict stock price while avoiding the data leakage and alleviating boundary problems. According to the experiments on two different datasets, our model mWDN-LSTM outperforms the well-known benchmarks such as CNN-LSTM on the same experimental setup and demonstrates the effectiveness of utilizing cyclic patterns in stock price prediction.
Background
This study aimed to investigate the occurrence and predictive factors of restenosis in coronary heart disease (CHD) patients underwent percutaneous coronary intervention (PCI) with ...sirolimus-eluting stent (SES).
Methods
Demographic data, clinical features, and laboratory tests of 398 CHD patients underwent PCI with SES were retrospectively reviewed. Coronary angiography was performed to evaluate coronary stenosis before PCI and in-stent restenosis at 1-year follow-up.
Results
There were 37 (9.3%) patients suffered restenosis, but 361 (90.7%) patients did not develop restenosis at 1-year follow-up. Demographic characteristic (age), cardiovascular risk factors (hypertension and hyperuricemia), biochemical indexes (fasting blood-glucose, total cholesterol, low density lipoprotein cholesterol (LDL-C) and high-sensitivity C-reactive protein (HsCRP)), cardiac function index (cardiac troponin I), lesion features (multivessel artery lesions, target lesion at left circumflex artery (LCX), two target lesions and length of target lesion), and operation procedure (length of stent) were correlated with higher restenosis risk. Moreover, age, hypertension, diabetes mellitus, LDL-C, HsCRP, and target lesion at LCX were independent predictive factors for raised restenosis risk. Based on these independent predictive factors, we established a restenosis risk prediction model, and receiver-operating characteristic curves displayed that this model exhibited an excellent predictive value for higher restenosis risk (areas under the curve 0.953 (95% CI 0.926–0.981)).
Conclusion
Our findings provide a new insight into the prediction for restenosis in CHD patients underwent PCI with SES.
The present study aimed to assess the correlation of fibroblast growth factor (FGF)-23 expression with clinical characteristics, then further explore its value in predicting 2-year in-stent ...restenosis (ISR) risk in coronary heart disease (CHD) patients underwent percutaneous coronary intervention (PCI) with drug-eluting stent (DES).
In this prospective, single-center, observational study, totally 214 CHD patients treated by PCI with DES were consecutively recruited, and peripheral blood samples were collected prior to PCI with DES for serum samples isolation. Following, FGF-23 level in the serum samples was detected via enzyme linked-immuno-sorbent Assay. The follow-up coronary angiography was performed at 1 year and 2 years after PCI or if suspected ISR symptoms occurred.
FGF-23 was positively correlated with fasting blood-glucose, insulin resistance, serum creatinine, serum uric acid, LDL-C, high-sensitivity C-reactive protein, cardiac troponin I and N-terminal-proB-type natriuretic peptide, while was negatively associated with HDL-C and left ventricular ejection fraction (all P < 0.01). Furthermore, FGF-23 was positively correlated with hypercholesteremia, hyperuricemia and family history of CAD (all P < 0.05). However, it did not correlate with other chronic complications, biochemical indexes, lesion features or PCI parameters (all P > 0.05). Moreover, FGF-23 level was higher in 2-year ISR patients (n = 38) compared to 2-year non-ISR patients (n = 176) (P < 0.001), and receiver operating characteristic curve indicated that FGF-23 was of good value in predicting 2-year ISR risk (AUC 0.828, 95% CI 0.761-0.896).
FGF-23 correlates with endocrine and metabolism dysregulation, worse cardiac and renal function, inflammation level, stenosis degree of target lesion, and serves as an independent risk factor for 2-year ISR risk in CHD patients underwent PCI with DES.
Contrast-induced nephropathy limits the outcomes of percutaneous coronary intervention (PCI). The present study compared the protective effects of different statin doses on renal function. A total of ...228 patients with acute coronary syndrome undergoing selective PCI were randomly divided into simvastatin 20-mg group (S20, n = 115) and simvastatin 80-mg group (S80, n = 113). Serum creatinine was measured at admission, the day of PCI, and 24 and 48 hours after PCI. The creatinine clearance was calculated using the Cochcroft-Gault formula. High-sensitive C-reactive protein, P-selectin, and intercellular adhesion molecule-1 were also measured before and after the procedure. Contrast-induced nephropathy was defined as a postprocedure increase in serum creatinine of ≥0.5 mg/dl or >25% from baseline. The serum creatinine significantly increased after PCI, with the peak value occurring at 24 hours, and then began to decrease. At 48 hours, the serum creatinine had decreased to the baseline level in the S80 group, but it had failed to do so in the S20 group. At 24 and 48 hours after PCI, the serum creatinine was lower in the S80 group than in the S20 group (p <0.05 and p <0.001, respectively). The creatinine clearance significantly decreased after PCI, with the lowest value occurring at 24 hours, and then it began to increase. In the S80 group, the creatinine clearance recovered to baseline level at 48 hours, but it failed to do so in the S20 group. The creatinine clearance was greater at 24 and 48 hours in the S80 group than that in the S20 group. Although the procedure caused a significant increase in high-sensitive C-reactive protein, P-selectin, and intercellular adhesion molecule-1 levels, the value was lower in the S80 group than in the S20 group (p <0.001). In conclusion, pretreatment with simvastatin 80 mg before PCI could further decrease the occurrence of contrast-induced nephropathy compared with simvastatin 20 mg. This benefit was associated with the lowering of high-sensitive C-reactive protein, P-selectin, and intercellular adhesion molecule-1 levels.
Prourokinase is a single-chain plasminogen activator presenting with fewer hemorrhagic complications and reduced reocclusion rate compared with the conventional fibrinolytic agents in patients with ...coronary artery disease. However, prourokinase intracoronary injection during PCI for treating patients with ST-segment elevation myocardial infarction (STEMI) is rarely investigated. Therefore, this study aimed to evaluate the efficacy and safety of intracoronary prourokinase during the percutaneous coronary intervention (PCI) in treating STEMI patients.
Fifty STEMI patients who underwent primary PCI were consecutively enrolled and randomly assigned to intracoronary prourokinase group (N = 25) or control group (N = 25). During the primary PCI procedure, patients in the intracoronary prourokinase group received 10 ml prourokinase injection, while patients in control group received 10 ml saline injection as control. The primary endpoints were coronary physiological indexes, the secondary endpoints were angiographic assessments, myocardial infarct size/reperfusion assessment, cardiac function evaluations, major adverse coronary events (MACEs) and hemorrhagic complications. All patients were followed up for 3 months.
Post PCI, the index of microcirculatory resistance (IMR) was decreased in intracoronary prourokinase group than that in control group (34.56 ± 7.48 vs. 49.00 ± 8.98, P < 0.001), while no difference of coronary flow reserve (CFR) (2.01 ± 0.32 vs. 1.88 ± 0.23, P = 0.267) or fractional flow reserve (FFR) (0.89 ± 0.05 vs. 0.87 ± 0.04, P = 0.121) was found between the two groups. The thrombolysis in myocardial infarction myocardial perfusion grade (TMPG) (P = 0.024), peak values of creatine kinase (CK) (P = 0.028), CK isoenzyme-MB (CK-MB) (P = 0.016), cardiac troponin I (cTnI) (P = 0.032) and complete ST-segment resolution (STR) (P = 0.005) were better in intracoronary prourokinase group compared with control group. At 3-months post PCI, left ventricular ejection fraction (LVEF) and wall motion score index (WMSI) were higher, while left ventricular end-diastolic diameter (LVEDd) was lower in intracoronary prourokinase group compared with control group (all P < 0.05). There was no difference in hemorrhagic complication or total MACE between the two groups.
Intracoronary prourokinase during PCI is more efficient and equally tolerant compared with PCI alone in treating STEMI patients.
Chinese Clinical Trial Registry ChiCTR1800016207 . Prospectively registered.
Urban crowd flow prediction is an important task for transportation systems and public safety. While graph convolutional networks (GCNs) have been widely adopted for this task, existing GCN-based ...methods still face challenges. Firstly, they employ fixed receptive fields, failing to account for urban region heterogeneity where different functional zones interact distinctly with their surroundings. Secondly, they lack mechanisms to adaptively adjust spatial receptive fields based on temporal dynamics, which limits prediction performance. To address these limitations, we propose an Adaptive Receptive Field Graph Convolutional Network (ARFGCN) for urban crowd flow prediction. ARFGCN allows each region to independently determine its receptive field size, adaptively adjusted and learned in an end-to-end manner during training, enhancing model prediction performance. It comprises a time-aware adaptive receptive field (TARF) gating mechanism, a stacked 3DGCN, and a prediction layer. The TARF aims to leverage gating in neural networks to adapt receptive fields based on temporal dynamics, enabling the predictive network to adapt to urban regional heterogeneity. The TARF can be easily integrated into the stacked 3DGCN, enhancing the prediction. Experimental results demonstrate ARFGCN’s effectiveness compared to other methods.
Stance detection is a crucial task in natural language processing and social computing, focusing on classifying expressed attitudes towards specific targets based on the input text. Conventional ...methods predominantly view stance detection as a task of target-oriented, sentence-level text classification. On popular social media platforms like Twitter, users often express their opinions through hashtags in addition to textual content within tweets. However, current methods primarily treat hashtags as data retrieval labels, neglecting to effectively utilize the semantic information they carry. In this paper, we propose a large language model knowledge-enhanced stance detection framework (LKESD) for stance detection. LKESD contains three main components: an instruction-prompted background knowledge acquisition module (IPBKA) that retrieves background knowledge of hashtags by providing handcrafted prompts to large language models (LLMs); a graph convolutional feature-enhancement module (GCFEM) is designed to extract the semantic representations of words that frequently co-occur with hashtags in the dataset by leveraging textual associations; an a knowledge fusion network (KFN) is proposed to selectively integrate graph representations and LLM features using a prompt-tuning framework. Extensive experimental results on three benchmark datasets demonstrate that our LKESD method outperforms 2.7% on all setups over compared methods, validating its effectiveness in stance detection tasks.