Prediction and Feature Importance of Earth Pressure in Shields Using Machine Learning Algorithms

Hongyu Huang,Lipeng Liu,Ruilang Cao,Yuxin Cao 대한토목학회 2023 KSCE Journal of Civil Engineering Vol.27 No.2

To reduce subjectivity and uncertainty when maintaining suitable earth pressure in earth pressure shields that can prevent heave or collapse, many prediction models using machine learning algorithms were proposed, but little research into the effects of other parameters on earth pressure has been undertaken, and soil conditioning parameters are always ignored. To establish a model with thorough parameters and probe into influences of other parameters, multiple machine learning algorithms were attempted. Given the accuracy, diversity and functions, random forest (RF), LightGBM and Attention-back-propagation neural network (Attention-BPNN) were further analyzed. Then, two RF models were compared in this research, one with soil conditioning parameters and the other without. Meanwhile, a case study was utilized to verify the reliability of the model. Finally, the feature importance of three models was compared and the variation rules of the most four important features were discussed by controlling variates. The results showed that soil conditioning parameters delivered a significant reduction in the prediction error. The case study demonstrated that the proposed model can satisfy engineering requirements. More earth pressure should give priority to increasing propulsion pressure, advance rate, and reducing foam air flow, rotational speed of screw conveyor, and vice versa.

GCNXSS: An Attack Detection Approach for Cross-Site Scripting Based on Graph Convolutional Networks

Hongyu Pan,Yong Fang,Cheng Huang,Wenbo Guo,Xuelin Wan 한국인터넷정보학회 2022 KSII Transactions on Internet and Information Syst Vol.16 No.12

Since machine learning was introduced into cross-site scripting (XSS) attack detection, many researchers have conducted related studies and achieved significant results, such as saving time and labor costs by not maintaining a rule database, which is required by traditional XSS attack detection methods. However, this topic came across some problems, such as poor generalization ability, significant false negative rate (FNR) and false positive rate (FPR). Moreover, the automatic clustering property of graph convolutional networks (GCN) has attracted the attention of researchers. In the field of natural language process (NLP), the results of graph embedding based on GCN are automatically clustered in space without any training, which means that text data can be classified just by the embedding process based on GCN. Previously, other methods required training with the help of labeled data after embedding to complete data classification. With the help of the GCN auto-clustering feature and labeled data, this research proposes an approach to detect XSS attacks (called GCNXSS) to mine the dependencies between the units that constitute an XSS payload. First, GCNXSS transforms a URL into a word homogeneous graph based on word co-occurrence relationships. Then, GCNXSS inputs the graph into the GCN model for graph embedding and gets the classification results. Experimental results show that GCNXSS achieved successful results with accuracy, precision, recall, F1-score, FNR, FPR, and predicted time scores of 99.97%, 99.75%, 99.97%, 99.86%, 0.03%, 0.03%, and 0.0461ms. Compared with existing methods, GCNXSS has a lower FNR and FPR with stronger generalization ability.

Preparation and Effect Analysis of {10–12} Extension Twins at the Edge of AZ31 Magnesium Alloy Plate

Zhiquan Huang,Chuanlu Qi,Yanchun Zhu,Guowei Yang,Hongyu Lai,Jinchao Zou 대한금속·재료학회 2021 METALS AND MATERIALS International Vol.27 No.11

The edge samples of AZ31 magnesium alloy plate after hot rolling and annealing were compressed along the rolling direction(RD) under different deformation conditions. Then, the optimal conditions for the generation of {10–12} extension twinsand the microtexture evolution of edge samples were studied by using EBSD. The results show that the texture strength ofthe hot-rolled AZ31 magnesium alloy plate was significantly greater than that of the middle part. When the compressivestrain along the RD was small at the edge of the rolled piece at room temperature, a large number of extension twin layerswere generated. Grain matrix and twins coexisted. As the strain increased to 8.19%, the grains completely transformed intoextension twins. Along with the generation of extension twins, hard-oriented grains were consumed in large quantities andthe weakening effect of the basal texture was proportional to the volume fraction of the extension twins, while at high temperature(250 °C), the slip dominated the plastic deformation process.

Ginsenoside compound K protects against cerebral ischemia/ reperfusion injury via Mul1/Mfn2-mediated mitochondrial dynamics and bioenergy

Qingxia Huang,Jing Li,Jinjin Chen,Zepeng Zhang,Peng Xu,Hongyu Qi,Zhaoqiang Chen,Jiaqi Liu,Jing Lu,Mengqi Shi,Yibin Zhang,Ying Ma,Daqing Zhao,Xiangyan Li The Korean Society of Ginseng 2023 Journal of Ginseng Research Vol.47 No.3

Background: Ginsenoside compound K (CK), the main active metabolite in Panax ginseng, has shown good safety and bioavailability in clinical trials and exerts neuroprotective effects in cerebral ischemic stroke. However, its potential role in the prevention of cerebral ischemia/reperfusion (I/R) injury remains unclear. Our study aimed to investigate the molecular mechanism of ginsenoside CK against cerebral I/R injury. Methods: We used a combination of in vitro and in vivo models, including oxygen and glucose deprivation/reperfusion induced PC12 cell model and middle cerebral artery occlusion/reperfusion induced rat model, to mimic I/R injury. Intracellular oxygen consumption and extracellular acidification rate were analyzed by Seahorse multifunctional energy metabolism system; ATP production was detected by luciferase method. The number and size of mitochondria were analyzed by transmission electron microscopy and MitoTracker probe combined with confocal laser microscopy. The potential mechanisms of ginsenoside CK on mitochondrial dynamics and bioenergy were evaluated by RNA interference, pharmacological antagonism combined with co-immunoprecipitation analysis and phenotypic analysis. Results: Ginsenoside CK pretreatment could attenuate mitochondrial translocation of DRP1, mitophagy, mitochondrial apoptosis, and neuronal bioenergy imbalance against cerebral I/R injury in both in vitro and in vivo models. Our data also confirmed that ginsenoside CK administration could reduce the binding affinity of Mul1 and Mfn2 to inhibit the ubiquitination and degradation of Mfn2, thereby elevating the protein level of Mfn2 in cerebral I/R injury. Conclusion: These data provide evidence that ginsenoside CK may be a promising therapeutic agent against cerebral I/R injury via Mul1/Mfn2 mediated mitochondrial dynamics and bioenergy.

Preparation and Aroma Analysis of Chinese Traditional Fermented Flour Paste

Yuyu Zhang,Mingquan Huang,Hongyu Tian,Baoguo Sun,Jing Wang,Quanhong Li 한국식품과학회 2014 Food Science and Biotechnology Vol.23 No.1

Volatile compounds in Chinese fermented flourpaste were extracted using simultaneous distillation andextraction (SDE) and analyzed using gas chromatographymassspectrometry (GC-MS) with DB-5 and DB-WAXcapillary columns. A total of 84 volatile compounds wereidentified, including 8 aldehydes, 19 esters, 14 acids, 17hydrocarbons, 7 heterocycles, and 19 other trace compounds. The major volatiles included furfural, 5-methyl-2-phenyl-2-hexenal, 4-ethylguaiacol, 2-phenylacetaldehyde, ethylhexadecanoate, isovaleraldehyde, palmitic acid, and 5-methylfurfural. Aroma compounds were investigated usinggas chromatography-olfactometry (GC-O) and aroma extractdilution analysis (AEDA). A total of 27 olfactory regionswere exposed and 18 aroma extracts were located. Altogether,6 aroma compounds identified using GC-O showed higherflavor dilution factors (FD≥32), including isovaleraldehyde,furfural, pentanoic acid, 2-acetylfuran, 1-octen-3-ol, and 2-phenylacetaldehyde. Compared with soybean sauce, fermentedflour paste has more esters and aldehydes, which contributeto the desired fruity, caramel, sweet, and roasted odors.

A Data Volume Aware Ant Colony Optimization Approach for Geographical Knowledge Cloud Service Composition

Xiaozhu Wu,Chongcheng Chen,Hongyu Huang 보안공학연구지원센터 2016 International Journal of Grid and Distributed Comp Vol.9 No.6

Geographical knowledge cloud service is a typical online service that provides big spatial data analysis with the function of knowledge discovery or decision-making. The composition of geographical knowledge cloud service imposes stricter requirements for better overall QoS and execution efficiency of the service chain. In this paper, we present a data volume aware ant colony optimization approach called DVA-MOACO algorithm for geographical knowledge cloud service composition. Our algorithm utilizes a multi-index service quality evaluation model, and improves the transition probability while considering the data transfer cost and other QoS constraints simultaneously when ant finding path. Our algorithm could reach the Pareto near optimal solution rapidly with better QoS performance and lower data transfer cost from numerous candidate solutions.

Study on constitutive model of 05Cr17Ni4Cu4Nb stainless steel based on quasi-static tensile test

Linlin Zhu,Xiongrong Huang,Hongyu Liu 대한기계학회 2022 JOURNAL OF MECHANICAL SCIENCE AND TECHNOLOGY Vol.36 No.6

In order to obtain a constitutive model that can accurately describe the stress flow law of 05Cr17Ni4Cu4Nb stainless steel in extrusion forming process of spherical plain bearings, the quasi-static tensile tests of this material were carried out at room temperature in this paper, the engineering stress-strain curves of the material were obtained. Six constitutive equations were obtained by fitting the stress-strain data between the yield point and the necking point, the errors of the six models with the experimental data were analyzed. The results indicate that the Swift-Voce constitutive model had the highest correlation coefficient, the smallest mean square deviation with the experimental data and the best prediction accuracy at pre-necking and post-necking stage. Therefore, the Swift-Voce constitutive model was selected as the constitutive model of 05Cr17Ni4Cu4Nb stainless steel with large deformation at room temperature, and the finite element simulation technology was used to verify the Swift-Voce constitutive model. Once again, the simulation results show that, the stress-strain curve at prenecking stage of the material is in good agreement with the test data. This constitutive model established in this paper can accurately describe the stress flow behavior of the material at room temperature.

Dose-Dense Rituximab-CHOP versus Standard Rituximab-CHOP in Newly Diagnosed Chinese Patients with Diffuse Large B-Cell Lymphoma: A Randomized, Multicenter, Open-Label Phase 3 Trial

Xueying Li,He Huang,Bing Xu,Hongqiang Guo,Yingcheng Lin,Sheng Ye,Jiqun Yi,Wenyu Li,Xiangyuan Wu,Wei Wang,Hongyu Zhang,Derong Xie,Jiewen Peng,Yabing Cao,Xingxiang Pu,Chengcheng Guo,Huangming Hong,Zhao 대한암학회 2019 Cancer Research and Treatment Vol.51 No.3

Purpose Rituximab with cyclophosphamide, doxorubicin, vincristine, and prednisone administered every 3 weeks (R-CHOP-21) is the standard care for diffuse large B-cell lymphoma (DLBCL). It is unknown whether the dose-dense R-CHOP (R-CHOP-14) could improve the outcome of the disease in Asian population. Materials and Methods Newly diagnosed DLBCL patients were centrally, randomly assigned (1:1) to receive R-CHOP- 14 or R-CHOP-21. R-CHOP-14 was administered every 2 weeks, and R-CHOP-21 was administered every 3 weeks. Primary end point was disease-free survival (DFS). Secondary end points included overall survival (OS), progression-free survival (PFS), response rate and toxicities. Results Seven hundred and two patients were randomly assigned to receive R-CHOP-14 (n=349) or R-CHOP-21 (n=353). With a median follow-up of 45.6 months, the two groups did not differ significantly in 3-year DFS (79.6% for R-CHOP-14 vs. 83.2% for R-CHOP-21, p=0.311), 3-year OS (77.5% for R-CHOP-14 vs. 77.6% for R-CHOP-21, p=0.903), or 3-year PFS (63.2% for R-CHOP-14 vs. 66.1% for R-CHOP-21, p=0.447). Patients with an International Prognostic Index (IPI) score ! 2 had a poorer prognosis compared to those with an IPI score < 2. Grade 3/4 hematologic and non-hematologic toxicities were manageable and similar between R-CHOP-14 and R-CHOP-21. Conclusion R-CHOP-14 did not improve the outcome of DLBCL compared to R-CHOP-21 in Asian population. With manageable and similar toxicities, both of the two regimens were suitable for Asian DLBCL patients. For high-risk patients with IPI ! 2, new combination regimens based on R-CHOP deserve further investigation to improve efficacy.

Automatic assessment of post-earthquake buildings based on multi-task deep learning with auxiliary tasks

Huamei Zhu,Zhihang Li,Mengqi Huang,Pengxuan Ji,Hongyu Huang,Qianbing Zhang 국제구조공학회 2023 Smart Structures and Systems, An International Jou Vol.31 No.4

Post-earthquake building condition assessment is crucial for subsequent rescue and remediation and can be automated by emerging computer vision and deep learning technologies. This study is based on an endeavour for the 2nd International Competition of Structural Health Monitoring (IC-SHM 2021). The task package includes five image segmentation objectives - defects (crack/spall/rebar exposure), structural component, and damage state. The structural component and damage state tasks are identified as the priority that can form actionable decisions. A multi-task Convolutional Neural Network (CNN) is proposed to conduct the two major tasks simultaneously. The rest 3 sub-tasks (spall/crack/rebar exposure) were incorporated as auxiliary tasks. By synchronously learning defect information (spall/crack/rebar exposure), the multi-task CNN model outperforms the counterpart single-task models in recognizing structural components and estimating damage states. Particularly, the pixel-level damage state estimation witnesses a mIoU (mean intersection over union) improvement from 0.5855 to 0.6374. For the defect detection tasks, rebar exposure is omitted due to the extremely biased sample distribution. The segmentations of crack and spall are automated by single-task U-Net but with extra efforts to resample the provided data. The segmentation of small objects (spall and crack) benefits from the resampling method, with a substantial IoU increment of nearly 10%.

Genotypic Characteristics of Haemophilus influenzae Isolates from Pediatric Pneumonia Patients in Chengdu City, Sichuan, China

Tian Guozhong,Zhang Li,Li Machao,Wang Xiaolei,Zheng Yuhong,Li Xiaojing,Huang Cheng,Li Xuechun,Xie Yongqiong,Xu Li,Ren Hongyu,Shao Zhujun 한국미생물학회 2009 The journal of microbiology Vol.47 No.4

Two hundred and seventy-three Haemophilus influenzae strains isolated from pediatric pneumonia patients in China were studied. We used Multilocus Sequence Typing (MLST) to analyze genotypic characteristics. All strains were biotyped and serotyped. Relatedness and patterns of genes among isolates were determined by the analysis of MLST and eBURST. H. influenzae primarily causes acute pneumonia in children under 1 year old. Nontypeable H. influenzae was responsible for most cases of pediatric pneumonia. All 273 strains were classified into eight biotypes. They mostly belonged to the I, II, and III biotypes (17.6%, 43.6%, and 22.7%, respectively). 62 strains (22.7%) produced β-lactamase. We found 28 novel alleles. Fifty different STs were found by MLST, of which 39 were novel. These were ST477 through ST508 and ST521 through ST527. Group 17 and predicted founders 503 were new groups in this study. No STs correlated with strains from Korea, which is adjacent to China. The H. influenzae strains from China appeared to have heterogeneous ST types patterns which may be the reason no outbreaks or epidemics of H. influenzae infections have occurred in Chengdu city, Sichuan, China.