Main Dataset Research Articles

Effect of noises removal and spatial resolutions of Digital Surface Model (DSM) in flood inundation model

Digital Elevation Model (DEM) is an important component in flood inundation modelling as a part of flood risk analysis. However, Digital Surface Model (DSM) which still contains noises of artefacts is more common to be available than DEM. Then DSM misleads the inundation extent. This study will evaluate a measure to obtain a bare-earth surface (DEM) by developing and applying a workflow to remove the noises from DSM using simple tools available in the GIS computer software. The output of the method will then be used as input for flood inundation modelling of Semarang City in Indonesia. In addition to the main data set of TerraSAR-X DSM provided by Indonesia Geospatial Information Agency (Badan Informasi Geospasial or BIG), two sets of open source digital elevation data from Shuttle Radar Topography Mission (SRTM) and Advanced Spaceborne Thermal Emission (ASTER) are included in this study. The three data sets are processed and quantitatively compared. The results show that the applied noise removal method must be traded off with the lower resolution. Hence, the level of detail of particular flood inundation modelling will determine the required DEM resolution. Different study purposes will lead to different appropriate DEM resolutions to be used.

Petrogenesis of main group pallasite meteorites based on relationships among texture, mineralogy, and geochemistry

AbstractMain group pallasite meteorites are samples of a single early magmatic planetesimal, dominated by metal and olivine but containing accessory chromite, sulfide, phosphide, phosphates, and rare phosphoran olivine. They represent mixtures of core and mantle materials, but the environment of formation is poorly understood, with a quiescent core–mantle boundary, violent core–mantle mixture, or surface mixture all recently suggested. Here, we review main group pallasite data sets and petrologic characteristics, and present new observations on the low‐MnO pallasite Brahin that contains abundant fragmental olivine, but also rounded and angular olivine and potential evidence of sulfide–phosphide liquid immiscibility. A reassessment of the literature shows that low‐MnO and high‐FeO subgroups preferentially host rounded olivine and low‐temperature P2O5‐rich phases such as the Mg‐phosphate farringtonite and phosphoran olivine. These phases form after metal and silicate reservoirs back‐react during decreasing temperature after initial separation, resulting in oxidation of phosphorus and chromium. Farringtonite and phosphoran olivine have not been found in the common subgroup PMG, which are mechanical mixtures of olivine, chromite with moderate Al2O3 contents, primitive solid metal, and evolved liquid metal. Lower concentrations of Mn in olivine of the low‐MnO PMG subgroup, and high concentrations of Mn in low‐Al2O3 chromites, trace the development and escape of sulfide‐rich melt in pallasites and the partially chalcophile behavior for Mn in this environment. Pallasites with rounded olivine indicate that the core–mantle boundary of their planetesimal may not be a simple interface but rather a volume in which interactions between metal, silicate, and other components occur.

Synchrotron radiation imaging revealing the sub-micron structure of the auditory ossicles

PurposeSynchrotron-based X-ray Phase Contrast Imaging (SR X–PCI) allows, thanks to a highly coherent and powerful X-ray beam, the imaging of surface and cross-sectional tissue properties with high absorption-contrast. The objective of this study is to investigate the sub-micron structure of the ossicular chain. The understanding of its morphological properties at sub-micron scale will help to refine the understanding of its structural properties. The investigation of intact, non-decalcified and unstained ossicular bones allows to study the spatial relationship between surface properties, internal structure and tomographical slides. Main resultsThe tomography datasets with a pixel size of 0.65 μm were reconstructed and 3D volume rendering models of all specimens were analyzed. Based on surface models, the surfaces of the articulations, the insertion of the tensor tympani and stapedial muscle tendons and the nutritional foramina, where the vessels penetrate the ossicles, were visualized. Moreover, a branched network of inner channels could be represented and its connection to the nutritional foramen was demonstrated. Looking at the tomographic structure of the three ossicles a mineralization pattern for every auditory bone was described, indicating a considerable variation throughout the bones. ConclusionsThis study investigates the submicron-structure of the auditory ossicles at a pixel size of 0.65 μm, which is to the best of our knowledge the highest resolution reported in the investigation of the human auditory system so far. The provided data helps in the further understanding of the anatomical conformation of the ossicular chain.

Acumulación de horas frío para la producción de arándano en Nayarit, México

El clima es un factor determinante del rendimiento y calidad en sistemas de producción agrícola. El objetivo de este trabajo fue cuantificar las horas frío-acumulables en Nayarit y utilizar dicha información como indicador agroclimático en la regionalización para la producción de arándano (Vaccinium corymbosum L.). En la investigación se emplearon datos de la red de estaciones agrometeorológicas de Nayarit, la cual se compone de 38 estaciones automatizadas que colectaron datos quinceminutales de 11 variables. Con registros de temperatura mínima se calcularon las horas frío, las cuales se organizaron por mes y acumulación por año. Mediante un sistema de información geográfica se representó cartográficamente la distribución de las horas frío anual (HFA), con interpolación y organización de los resultados en cuatro fases (<299, 300-399, 400-499 y 500-699). Para cada clase de HFA, se cuantificó la superficie en cada uno de los 20 municipios. Finalmente se evaluó la precisión del modelo de interpolación contra una base de datos ajena al conjunto de datos principal. Los resultados muestran cómo solamente en el 27.4% de la superficie estatal se acumulan más de 300 HFA que es el requerimiento mínimo para arándano. Las condiciones térmicas para arándano se ubican principalmente en el centro y sur del estado, resaltando los municipios de Compostela, Xalisco, Tepic, Santa María del Oro, Jala, La Yesca e Ixtlán del Río que representan el 81.6% de la superficie apta. Los resultados son un precedente sobre la acumulación de horas de frío en Nayarit y funge como instrumento de planeación para tomadores de decisión, técnicos y productores interesados en el cultivo de arándano.

Recognition of words from brain-generated signals of speech-impaired people: Application of autoencoders as a neural Turing machine controller in deep neural networks

There is an essential requirement to support people with speech and communication disabilities. A brain–computer interface using electroencephalography (EEG) is applied to satisfy this requirement. A number of research studies to recognize brain signals using machine learning and deep neural networks (DNNs) have been performed to increase the brain signal detection rate, yet there are several defects and limitations in the techniques. Among them is the use in specific circumstances of machine learning. On the one hand, DNNs extract the features well and automatically. On the other hand, their use results in overfitting and vanishing problems. Consequently, in this research, a deep network is designed on the basis of an autoencoder neural Turing machine (DN-AE-NTM) to resolve the problems by the use of NTM external memory. In addition, the DN-AE-NTM copes with all kinds of signals with high detection rates. The data were collected by P300 EEG devices from several individuals under the same conditions. During the test, each individual was requested to skim images with one to six labels and focus on only one of the images. Not to focus on some images is analogous to producing unimportant information in the individual’s brain, which provides unfamiliar signals. Besides the main P300 EEG dataset, EEG recordings of individuals with alcoholism and control individuals and the EEGMMIDB, MNIST, and ORL datasets were implemented and tested. The proposed DN-AE-NTM method classifies data with an average detection rate of 97.5%, 95%, 98%, 99.4%, and 99.1%, respectively, in situations where the signals are noisy so that only 20% of the data are reliable and include useful information.

Long-term transient groundwater pressure and deep infiltration in Alpine mountain slopes (Poschiavo Valley, Switzerland)

Bedrock aquifers in alpine catchments are important regional sources of freshwater. Data regarding bedrock groundwater-recharge processes are scarce and governing dynamics are poorly understood. The main datasets used to constrain regional groundwater recharge and flow, so far, have been based on indirect methods (environmental isotopes, river discharge rates, tunnel inflows). Here, a unique dataset is presented of long-term pore-water-pressure measurements from five deep boreholes situated in the upper reaches of a mountain slope at 1,500–2,300 m above sea level. In addition to multilevel pore pressure records, a detailed analysis of the hydrogeological conditions in the Alpine catchment is provided, along with the results from investigations of groundwater recharge mechanisms in response to variations in climatic conditions and hydraulic rock-mass properties. The recorded pressure data show annual pressure variations with amplitudes of 5–45 m and responses within a few days to summer rainstorms in the available depth range of 45–277 m below ground surface. One-dimensional analytical pore-pressure diffusion models and numerical infiltration models were applied to investigate pore-pressure dynamics and water-table variations. The model results reproduced the following parameters for the uppermost 100-m-thick layer: the observed amplitudes, rates and delays of pressure increase with porosities of 0.05–0.1%, specific storage of 5E-5 to 5E-7 m-1, and hydraulic diffusivities of 1E-1 to 1E-3 m2/s. Boreholes located in high-diffusivity areas (strongly fractured bedrock below coarse slope debris) had the strongest pressure variations and were most sensitive to weather conditions.

The Role of Science Knowledge Application in Improving Engineering Problem Solving Skills

This study presents how two types of integrated science and engineering lessons affect students’ engineering problem solving skills and their perceptions of engineering. In total, 146 middle school students participated in this study. Eighty-six students participated in the Type I lesson (complete engineering design lesson with a science knowledge application) and 60 students participated in the Type II lesson (engineering design without a science knowledge application). Two main datasets, (1) students’ Creative Engineering Problem Solving Propensity (CEPSP) measurement scores and (2) open-ended survey questions about students’ perceptions of engineering, were collected before and after the lessons. The results of this study show that after participating in the Type I lesson, students’ CEPSP scores significantly increased, whereas the CEPSP scores of the students who participated in the Type II lesson did not increase significantly. In addition, students who participated in the Type I lesson perceived engineering and the engineering integrated science lesson differently compared to the students who participated in the Type II lesson. The results of this study show that engineering integrated science, technology, engineering & mathematics (STEM) lessons should include a complete engineering design and a science knowledge application to improve students’ engineering problem solving skills.

Psychological and social determinants of physical activity from diagnosis to remission among French cancer patients (PERTINENCE): protocol for a mixed-method study

BackgroundMany effective physical activity (PA) interventions have focused on individual factors or a single theoretical model, limiting our understanding of the determinants of PA practice and their interactions in the cancer trajectory. The present mixed-method study aims to capture social and psychological determinants of PA practice from diagnosis to remission among cancer patients, and to identify key levers for PA practice.Methods/designA nested sequential mixed-method design QUAN (QUAL+QUAL) will be used, with qualitative studies embedded in the quantitative study to broaden our understanding of the determinants of PA practice. The design is sequential, since qualitative data on medical staff will be collected before patient inclusion (Phase 1), followed by quantitative patient data collection lasting one year (Phase 2) and a final qualitative data collection one year after inclusion (Phase 3). Phase 1 will be a case study in the two hospitals involved in the study, exploring knowledge of and support for PA practice among medical staff. Through interviews and documental analyses, the PA support dynamic will be evaluated with regard to PA prescription. Phase 2 will be a one-year observational study among 693 cancer patients. Quantitative medical, social, dispositional and psychological data, PA practices and preferences, will be collected at diagnosis, and six months and one year thereafter. Phase 3 will be a retrospective study, evaluating societal and policy factors, as well as unexpected factors playing a role in PA levels and preferences among cancer patients. For this phase thirty patients will be identified six months after inclusion on the basis of their PA profiles. Quantitative data will provide the main dataset, whilst qualitative data will complete the picture, enabling determinants of PA practice and their interactions to be captured throughout the cancer trajectory.DiscussionThe present study aims to identify key levers and typical trajectories for PA practice among cancer patients, adapted to different times in the course of cancer and taking into account “what works”, “for whom”, “where” and “how”. The challenge is the tailoring of PA interventions to patients at different times in their cancer trajectory, and the implication of medical staff support.Trial registrationClinical Trial NCT03919149, 18 April 2019. Prospectively registered.

Manifesting multidimensional creativity in a technology‐mediated online TESOL practicum course

This qualitative case study examined how creativity was manifested in a technology‐mediated online TESOL practicum course with six in‐service teachers in a midwestern U.S. teacher education program. The study defines creativity as ability to explore, adapt, take risks, collaborate, communicate, reflect on, and innovate with novel, unpredictable, and appropriate resources. To examine a research question of how creativity was manifested in an online practicum course, two main data sets were analyzed: VoiceThread projects (talks and texts) and final papers. Findings are presented under each of three themes of creativity in action: creativity as collaborative critical reflection, creativity transformed as new pedagogy, and creativity mediated by technology. First, creativity was exercised as collaborative participation by which practicum students shared ideas, suggestions, and resources through a high‐level reflexivity among peers through guided feedback. Second, creativity was translated as new linguistically and culturally responsive teaching‐oriented pedagogical plans and practices. Finally, innovative use of the instructional technology—teaching videos and VoiceThread—was the key to enhancing teachers’ critical and transformative view of their own teaching and to increasing opportunities to learn in a safe and collaborative communal space.

Second-generation antipsychotic drugs and short-term somatic serious adverse events: a systematic review and meta-analysis

Antipsychotic drugs might cause acutely occurring, serious side-effects and thus contribute to the increased physical morbidity and mortality observed in patients with severe mental health disorders. We examined this hypothesis by doing a meta-analysis of International Conference on Harmonisation-Good Clinical Practice-defined serious adverse events occurring in placebo-controlled trials of antipsychotics. For this systematic review and meta-analysis, we included randomised controlled trials (RCTs) comparing second-generation antipsychotics with placebo. We searched MEDLINE, Embase, Cochrane CENTRAL, BIOSIS, PsycINFO, PubMed, ClinicalTrials.gov, and WHO International Clinical Trials Registry Platform for trials published in any language from database inception up until Jan 27, 2017. Trials were included without limitations in population (diagnostic category, age, sex, ethnicity), dosing regimen, blinding status, duration, or publication year. Only psychological studies lasting less than 1 day and trials done in mainland China were excluded. We contacted pharmaceutical companies, drug regulatory authorities, and study investigators for additional data. The primary outcome was the number of patients with at least one somatic serious adverse event. We estimated minimum and maximum numbers of patients with the outcome in each study group and synthesised the results with odds ratios (ORs) in a common-effects meta-analysis. This study is registered with PROSPERO, number CRD42016033930. We identified 597 RCTs, comprising 108 664 participants, that met the inclusion criteria. 314 trials (67 642 participants) with details on individual serious adverse events available constituted the main dataset for meta-analysis. 88% of these were 13 weeks (approximately 3 months) or shorter in duration (median 6 weeks, IQR 4-9). At least one somatic serious adverse event occurred in 698 (1·63%) to 862 (2·02%) of 42 600 patients on antipsychotics, and in 343 (1·37%) to 419 (1·67%) of 25 042 patients on placebo. The odds ratios (ORs) were 1·24 (95% CI 1·08-1·42) and 1·24 (1·10-1·41) based on the minimum and maximum estimate, respectively. In predefined subgroup analyses we found evidence suggesting a larger effect in older patients (>65 years; OR 1·56, 95% CI 1·22-1·98; 1·58, 1·25-1·99) as compared with adults (18-65 years; 1·09, 0·91-1·29; 1·10, 0·95-1·28); likewise in children or adolescents (<18 years) although the evidence was more uncertain (1·49, 0·81-2·75; 1·54, 0·85-2·77). Of 597 included RCTs, 30 (5%), 358 (60%), and 209 (35%) were rated at high, moderate, or low risk of bias, respectively. τ2 was zero for both analyses of the primary outcome (minimum estimate, maximum estimate). A Bayesian sensitivity analysis using external information on heterogeneity gave similar results. We found evidence that antipsychotics cause short-term somatic serious adverse events on top of somatic serious adverse events occurring independent of treatment. This effect appears to be mainly driven by results in older patients. Hence, clinicians should be aware that antipsychotics are potentially toxic, particularly when treating patients sharing risk factors with the older population. German Ministry of Education and Research.

Automated deep learning method for whole‐breast segmentation in diffusion‐weighted breast MRI

Diffusion-weighted imaging (DWI) in MRI plays an increasingly important role in diagnostic applications and developing imaging biomarkers. Automated whole-breast segmentation is an important yet challenging step for quantitative breast imaging analysis. While methods have been developed on dynamic contrast-enhanced (DCE) MRI, automatic whole-breast segmentation in breast DWI MRI is still underdeveloped. To develop a deep/transfer learning-based segmentation approach for DWI MRI scans and conduct an extensive study assessment on four imaging datasets from both internal and external sources. Retrospective. In all, 98 patients (144 MRI scans; 11,035 slices) of four different breast MRI datasets from two different institutions. 1.5T scanners with DCE sequence (Dataset 1 and Dataset 2) and DWI sequence. A 3.0T scanner with one external DWI sequence. Deep learning models (UNet and SegNet) and transfer learning were used as segmentation approaches. The main DCE Dataset (4,251 2D slices from 39 patients) was used for pre-training and internal validation, and an unseen DCE Dataset (431 2D slices from 20 patients) was used as an independent test dataset for evaluating the pre-trained DCE models. The main DWI Dataset (6,343 2D slices from 75 MRI scans of 29 patients) was used for transfer learning and internal validation, and an unseen DWI Dataset (10 2D slices from 10 patients) was used for independent evaluation to the fine-tuned models for DWI segmentation. Manual segmentations by three radiologists (>10-year experience) were used to establish the ground truth for assessment. The segmentation performance was measured using the Dice Coefficient (DC) for the agreement between manual expert radiologist's segmentation and algorithm-generated segmentation. The mean value and standard deviation of the DCs were calculated to compare segmentation results from different deep learning models. For the segmentation on the DCE MRI, the average DC of the UNet was 0.92 (cross-validation on the main DCE dataset) and 0.87 (external evaluation on the unseen DCE dataset), both higher than the performance of the SegNet. When segmenting the DWI images by the fine-tuned models, the average DC of the UNet was 0.85 (cross-validation on the main DWI dataset) and 0.72 (external evaluation on the unseen DWI dataset), both outperforming the SegNet on the same datasets. The internal and independent tests show that the deep/transfer learning models can achieve promising segmentation effects validated on DWI data from different institutions and scanner types. Our proposed approach may provide an automated toolkit to help computer-aided quantitative analyses of breast DWI images. 3 Technical Efficacy: Stage 2 J. Magn. Reson. Imaging 2020;51:635-643.

First series of seafood datasets in ecoinvent: setting the pace for future development

Purpose: The number of life cycle assessment studies related to seafood has risen considerably in the past decade. Despite this proliferation, major life cycle inventory databases tend to lack information describing this sector. Hence, the main objectives of this study are to present the first effort to aggregate and standardize seafood-related datasets in the ecoinvent database and to explain the main data sources and methodological choices used in the building of the datasets. Methods: A list of the main datasets included in this first series is presented with a brief description of the underlying modelling approaches. Seafood capture, production and processing activities were modelled as the use phase of the required infrastructure. The full life cycle of infrastructure was considered, from construction, through use and maintenance to end-of-life. Results and discussion: Some of the most representative seafood industries in South America were modelled, namely Peruvian anchovy and hake fisheries, Andean trout, Brazilian tilapia and Peruvian fishmeal production, as well as the production of canned, frozen, cured and of a multi-ingredient fish-based product (fish sticks). Inventory data were found to be in line with those of seafood LCA literature and driven by the parameters widely known to be critical: fuel use intensity for fisheries, feed conversion ratio for aquaculture and energy intensity for seafood processing and reduction into fishmeal. The modelling approach was modular and intuitive, thus useful and reproducible by database users and data providers. Conclusions: The datasets created constitute a robust basis for the use of seafood-related data in international databases. It is expected that this work will stimulate further efforts by practitioners and data providers to model their inventory data into ecoinvent and other life cycle inventory databases.

Gait analysis and recognition prediction of the human skeleton based on migration learning

Gait recognition is a hot topic in the computing. Different gaits have different characteristics. This paper predicts whether a person in an image or video is running or walking by capturing the behavior of a character in an image or video. In this paper, we identify the gait by adopting the method of transfer learning and inception -V3 neural network. At the same time, the paper uses the HMDB - a large human motion database and UCF sports actions video action data as the main data set. At the end, this will help predict if the characters in either the picture or video make a running or walking motion. Results show a significant increase in object detection performance in comparison to existing algorithms with the use of transfer learning neural networks adapted for mobile use.

CNN-Based Dense Image Matching for Aerial Remote Sensing Images

Dense stereo matching plays a key role in 3D reconstruction. The capability of using deep learning in the stereo matching of remote sensing data is currently uncertain. This article investigated the application of deep learning–based stereo methods in aerial image series and proposed a deep learning–based multi-view dense matching framework. First, we applied three typical convolutional neural network models, MC-CNN, GC-Net, and DispNet, to aerial stereo pairs and compared the results with those of the SGM and a commercial software, SURE. Second, on different data sets, the generalization ability of each network is evaluated by using direct transfer learning with models pretrained on other data sets and by fine-tuning with a small number of target training data. Third, we present a deep learning–based multi-view dense matching framework where the multi-view geometry is introduced to further refine matching results. Three sets of aerial images as the main data sets and two open-source sets of street images as auxiliary data sets are used for testing. Experiments show that, first, the performance of deep learning–based stereo methods is slightly better than traditional methods. Second, both the GC-Net and the MC-CNN have demonstrated good generalization ability and can obtain satisfactory results on aerial images using a pretrained model on several available stereo benchmarks. Third, multi-view geometry constraints can further improve the performance of deep learning–based methods, which is better than that of the multi-view–based SGM and SURE.

Extended common spatial and temporal pattern (ECSTP): A semi-blind approach to extract features in ERP detection

Common spatial pattern (CSP) analysis and its extensions have been widely used as feature extraction approaches in the brain-computer interfaces (BCIs). However, most of the CSP-based approaches do not use any prior knowledge that might be available about the two conditions (classes) to be classified. Therefore, their applications are limited to datasets that contain enough variance information about the two conditions. For example, in some event-related potential (ERP) detection applications, such as P300 speller, the information is in the time domain but not in the variance of spatial components. To address this problem, first, we present a novel feature extraction method termed extended common spatial pattern (ECSP) analysis, which uses prior knowledge available from data to produce a broader range of features than that of conventional CSP analysis. Then, similarly, we introduce the extended common temporal pattern (ECTP) analysis. Finally, to exploit both spatial and temporal information, we propose extended common spatial and temporal pattern (ECSTP) analysis. We have used BCI competition III, dataset II as our main dataset to evaluate our proposed methods. In addition, we used two other datasets, namely BCI competition II, dataset IIb and BCI competition IV, dataset IIb, to further evaluate the performance of the proposed methods. In All the datasets, the proposed methods significantly outperform the conventional CSP, CTP, and CSTP methods. More specifically, ECSTP has the best performance among the proposed methods. Moreover, classification results show that the proposed methods are competitive with other state of the art methods applied to these datasets.

Motivation to English Academic Writing: Chinese Students’ Literacy Autobiography

This study looks at literacy as a site of power in which some forms of literacy practices have more power. Such power relations is even more complicated in English as a foreign language context, where school literacy, family literacy, and English language come into the picture. Many studies have explored different research methods to examine language learners’ voices in literacy practices, while literacy autobiography (LA) is not frequently used. LA is a reflective, first-person narrative of personal engagement in literacy practices. Through the LAs from a group of third-year university students in China, this qualitative case study examines motivation to in-school writing in general, their L1 and L2 in- and out-of-school writing experiences in particular. The main dataset includes 25 participants’ LAs; critical discourse analysis was used for data analysis. The findings suggest that their motivation to English academic writing is closely intertwined with their L1 academic writing, that it is the pedagogies employed in both L1 and L2 in-school writing that has affected their engagement, and that EFL students’ border-crossing experiences between both temporal and spatial spaces have shaped their engagement in writing. More important, demotivated students are not necessarily slow students in writing classes.

Gait-based Person Re-identification

The way people walk is a strong correlate of their identity. Several studies have shown that both humans and machines can recognize individuals just by their gait, given that proper measurements of the observed motion patterns are available. For surveillance applications, gait is also attractive, because it does not require active collaboration from users and is hard to fake. However, the acquisition of good-quality measures of a person’s motion patterns in unconstrained environments, (e.g., in person re-identification applications) has proved very challenging in practice. Existing technology (video cameras) suffer from changes in viewpoint, daylight, clothing, accessories, and other variations in the person’s appearance. Novel three-dimensional sensors are bringing new promises to the field, but still many research issues are open. This article presents a survey of the work done in gait analysis for re-identification in the past decade, looking at the main approaches, datasets, and evaluation methodologies. We identify several relevant dimensions of the problem and provide a taxonomic analysis of the current state of the art. Finally, we discuss the levels of performance achievable with the current technology and give a perspective of the most challenging and promising directions of research for the future.

The Cholecystectomy As A Day Case (CAAD) Score: A Validated Score of Preoperative Predictors of Successful Day-Case Cholecystectomy Using the CholeS Data Set

BackgroundDay-case surgery is associated with significant patient and cost benefits. However, only 43% of cholecystectomy patients are discharged home the same day. One hypothesis is day-case cholecystectomy rates, defined as patients discharged the same day as their operation, may be improved by better assessment of patients using standard preoperative variables.MethodsData were extracted from a prospectively collected data set of cholecystectomy patients from 166 UK and Irish hospitals (CholeS). Cholecystectomies performed as elective procedures were divided into main (75%) and validation (25%) data sets. Preoperative predictors were identified, and a risk score of failed day case was devised using multivariate logistic regression. Receiver operating curve analysis was used to validate the score in the validation data set.ResultsOf the 7426 elective cholecystectomies performed, 49% of these were discharged home the same day. Same-day discharge following cholecystectomy was less likely with older patients (OR 0.18, 95% CI 0.15–0.23), higher ASA scores (OR 0.19, 95% CI 0.15–0.23), complicated cholelithiasis (OR 0.38, 95% CI 0.31 to 0.48), male gender (OR 0.66, 95% CI 0.58–0.74), previous acute gallstone-related admissions (OR 0.54, 95% CI 0.48–0.60) and preoperative endoscopic intervention (OR 0.40, 95% CI 0.34–0.47). The CAAD score was developed using these variables. When applied to the validation subgroup, a CAAD score of ≤5 was associated with 80.8% successful day-case cholecystectomy compared with 19.2% associated with a CAAD score >5 (p < 0.001).ConclusionsThe CAAD score which utilises data readily available from clinic letters and electronic sources can predict same-day discharges following cholecystectomy.

Highly Accurate Machine Fault Diagnosis Using Deep Transfer Learning

We develop a novel deep learning framework to achieve highly accurate machine fault diagnosis using transfer learning to enable and accelerate the training of deep neural network. Compared with existing methods, the proposed method is faster to train and more accurate. First, original sensor data are converted to images by conducting a Wavelet transformation to obtain time-frequency distributions. Next, a pretrained network is used to extract lower level features. The labeled time-frequency images are then used to fine-tune the higher levels of the neural network architecture. This paper creates a machine fault diagnosis pipeline and experiments are carried out to verify the effectiveness and generalization of the pipeline on three main mechanical datasets including induction motors, gearboxes, and bearings with sizes of 6000, 9000, and 5000 time series samples, respectively. We achieve state-of-the-art results on each dataset, with most datasets showing test accuracy near 100%, and in the gearbox dataset, we achieve significant improvement from 94.8% to 99.64%. We created a repository including these datasets located at mlmechanics.ics.uci.edu.

A regression-tree multilayer-perceptron hybrid strategy for the prediction of ore crushing-plate lifetimes

Highly tensile manganese steel is in great demand owing to its high tensile strength under shock loads. All workpieces are produced through casting, because it is highly difficult to machine. The probabilistic aspects of its casting, its variable composition, and the different casting techniques must all be considered for the optimisation of its mechanical properties. A hybrid strategy is therefore proposed which combines decision trees and artificial neural networks (ANNs) for accurate and reliable prediction models for ore crushing plate lifetimes. The strategic blend of these two high-accuracy prediction models is used to generate simple decision trees which can reveal the main dataset features, thereby facilitating decision-making. Following a complexity analysis of a dataset with 450 different plates, the best model consisted of 9 different multilayer perceptrons, the inputs of which were only the Fe and Mn plate compositions. The model recorded a low root mean square error (RMSE) of only 0.0614 h for the lifetime of the plate: a very accurate result considering their varied lifetimes of between 746 and 6902 h in the dataset. Finally, the use of these models under real industrial conditions is presented in a heat map, namely a 2D representation of the main manufacturing process inputs with a colour scale which shows the predicted output, i.e. the expected lifetime of the manufactured plates. Thus, the hybrid strategy extracts core training dataset information in high-accuracy prediction models. This novel strategy merges the different capabilities of two families of machine-learning algorithms. It provides a high-accuracy industrial tool for the prediction of the full lifetime of highly tensile manganese steel plates. The results yielded a precision prediction of (RMSE of 0.061 h) for the full lifetime of (light, medium, and heavy) crusher plates manufactured with the three (experimental, classic, and highly efficient (new)) casting methods.