Suitable Datasets Research Articles

Topic prediction and knowledge discovery based on integrated topic modeling and deep neural networks approaches

Understanding the real-world short texts become an essential task in the recent research area. The document deduction analysis and latent coherent topic named as the important aspect of this process. Latent Dirichlet Allocation (LDA) and Probabilistic Latent Semantic Analysis (PLSA) are suggested to model huge information and documents. This type of contexts’ main problem is the information limitation, words relationship, sparsity, and knowledge extraction. The knowledge discovery and machine learning techniques integrated with topic modeling were proposed to overcome this issue. The knowledge discovery was applied based on the hidden information extraction to increase the suitable dataset for further analysis. The integration of machine learning techniques, Artificial Neural Network (ANN) and Long Short-Term (LSTM) are applied to anticipate topic movements. LSTM layers are fed with latent topic distribution learned from the pre-trained Latent Dirichlet Allocation (LDA) model. We demonstrate general information from different techniques applied in short text topic modeling. We proposed three categories based on Dirichlet multinomial mixture, global word co-occurrences, and self-aggregation using representative design and analysis of all categories’ performance in different tasks. Finally, the proposed system evaluates with state-of-art methods on real-world datasets, comprises them with long document topic modeling algorithms, and creates a classification framework that considers further knowledge and represents it in the machine learning pipeline.

COASTAL FLOODING RISK ASSESSMENT BY A NEURAL NETWORK APPROACH

An effective system of coastal flooding forecasting in the case of storm is essential to mitigate coastal risks for the population living in low-land coastal zones (less than 10 m above MSL). Nowadays, predictions of coastal flooding are usually carried out by adopting nested numerical models. However, the models adopted to obtain the data in the nearshore area require high computational costs, which are often too demanding and not viable for large scale forecasting. Data-driven models, such as Artificial Neural Networks (ANNs) can help to solve the problem as they can map complex nonlinear relationships between input and output variables once a suitable dataset of process realizations is available. In the present study a forecasting model for coastal flooding based on ANNs, in which the input data are the offshore wave characteristics from large scale model and the output results are the flooded areas, is proposed. These outputs provided a straightforward prediction of the area interested by coastal flooding during storms. Here an application of the model to assess the flooding risk in the village of Granelli, in the Southeast of Sicily (Italy) is presented.

Intelligent Character Recognition System Using Convolutional Neural Network

Computational Linguistics involves the techniques of Computer Science which play a vital role in recognizing written or printed characters such as numbers or letters to change them into a form that the computer can use it efficiently. Convolutional Neural Network differs from other approaches by extracting the features automatically. The proposed approach is capable of recognizing characters in a variety of challenging conditions using the Convolutional Neural Network, where traditional character recognition systems fail, notably in the presence of low resolution, substantial blur, low contrast, and other distortions. Intellectual Character Recognition System is an application that uses Convolutional Neural Network (CNN) to recognize the Tamil character dataset accurately developed by HP Labs India. The novelty of this system is that, it recognizes the characters of the Predominant Tamil language. With the help of suitable datasets consisting of the Tamil Scripts, the model is trained efficiently. This work has produced a training accuracy of 99.16% which is far better compared to the traditional approaches.

Flow-Data Gathering Using NetFlow Sensors for Fitting Malicious-Traffic Detection Models.

Advanced persistent threats (APTs) are a growing concern in cybersecurity. Many companies and governments have reported incidents related to these threats. Throughout the life cycle of an APT, one of the most commonly used techniques for gaining access is network attacks. Tools based on machine learning are effective in detecting these attacks. However, researchers usually have problems with finding suitable datasets for fitting their models. The problem is even harder when flow data are required. In this paper, we describe a framework to gather flow datasets using a NetFlow sensor. We also present the Docker-based framework for gathering netflow data (DOROTHEA), a Docker-based solution implementing the above framework. This tool aims to easily generate taggable network traffic to build suitable datasets for fitting classification models. In order to demonstrate that datasets gathered with DOROTHEA can be used for fitting classification models for malicious-traffic detection, several models were built using the model evaluator (MoEv), a general-purpose tool for training machine-learning algorithms. After carrying out the experiments, four models obtained detection rates higher than 93%, thus demonstrating the validity of the datasets gathered with the tool.

Improving the capability of integrated DInSAR and PSI approach for better detection, monitoring, and analysis of land surface deformation in underground mining environment

The study involves developing a framework of research methodology for enhancing the capability of spaceborne integrated DInSAR and PSI approach for better detection, monitoring, and analysis of land surface deformation in the underground mining environment. First component of the framework involves pre-feasibility assessment of the integrated approach followed by selecting suitable InSAR data sets considering three significant parameters, i.e., terrain, deformation, and system characteristics. The other essential components are preprocessing of the selected & acquired InSAR data for its quality check, the scheme for better detection as well as monitoring of mining subsidence trough(s) using Coherent Small Baseline-DInSAR Time Series (CSB-DTS), scheme for better in-depth monitoring of detected and undetected mining subsidence trough(s) using PSI, and result validation & analysis. Effectiveness of the proposed framework has been validated in the assessment of subsidence due to underground mining activities in one of the largest copper-ore producing belts in India, Khetri Copper Belt.

Predicting the ratings of Amazon products using Big Data

AbstractThis paper aims to apply several machine learning (ML) models to the massive dataset present in the area of e‐commerce from Amazon to analyze and predict ratings and to recommend products. For this purpose, we have used both traditional and Big Data algorithms. As the Amazon product review dataset is large, we present Big Data architecture suitable massive dataset for storing and computation, which is not possible with the traditional architecture. Furthermore, the dataset contains 15 attributes and has about 7 million records. With the dataset, we develop several models in Oracle Big Data and Azure Cloud Computing services to predict the review rating and recommendation for the items at Amazon. We present a comparative conclusion in terms of the accuracy as well as the efficiency with Spark ML—the Big Data architecture, and Azure ML—the traditional architecture.This article is categorized under: Fundamental Concepts of Data and Knowledge > Big Data Mining Technologies > Machine Learning Technologies > Prediction

An open toolbox for generating map of actively confirmed SARS-CoV-2 or COVID-19 cases in Vietnam

The recent outbreak of novel coronavirus, SARS-CoV-2 or COVID-19, discovered in late 2019, being continued to spread across regions worldwide, has resulted in 1,914,916 “confirmed” cases with up to 123,010 deaths, as in situation report –85 by World Health Organization (WHO). Most of the developed disease monitoring and tracking tools currently available only present the reported cases up to country-level and not detail down to provincial- or state-, city- level within the countries. This is insignificant for supporting activities in quickly reducing and preventing the spread of the disease within a certain country because further detail potential infectious locations are not provided for people to avoid traveling or passing by there. Thus, this work presents an open toolbox for generating map of actively “Confirmed” cases in a country, i.e., Vietnam, given a dataset containing their statuses and current locations, detail down to provincial-or state-, city-level. The newly released algorithm reduced approximately 24.41% of processing time of the preceding one. In addition, the algorithm can be easily extended for supporting other countries given suitable datasets.

How much spatial resolution do we need to model a local flood event? Benchmark testing based on UAV data from Biga River (Turkey)

Digital elevation models (DEMs) are important in flood modelling as they present surface information and obstacles giving the direction of water during a flood event. All hydraulic modelling software needs both channel and floodplain elevation data to simulate the water flow from upstream to downstream. Based on the accuracy and representative capacity of the surface, modelling results can be evaluated in terms of hazard and risk analysis. Developments in remote sensing and GIS technology provide many sources to freely obtain DEM data of different resolution and accuracy. However, access to and the processing of high-resolution data associated with large areas is difficult and time-consuming. This study, therefore, focuses on determining the availability of different resolutions and DEM sources in flood modelling by testing unmanned aerial vehicle (UAV) data. For this, freely available DEMs such as ALOS DEMs (12.5 and 30 m), EU-DEM (25 m), ASTER GDEM (30 m), SRTM (30 and 90 m), MERIT DEM (90 m) and TANDEM-X (90 m) were chosen to investigate part of Biga River (Turkey) which features a natural channel. Then, 1D flood modelling was examined using different DEM resolutions, Manning’s values and flood data of various frequencies (10, 100 and 1000 years). Results were tested using high-resolution DEM data (5 cm) gathered from a UAV. It was concluded that EU-DEM is the most suitable dataset in terms of flood extent, depth and velocity. However, depending on the boundary characteristics of the data, ALOS 12.5 DEM and MERIT DEM (90 m) data are the most suitable in their extent, depth and velocity for high-resolution and low-resolution flood modelling worldwide extent.

BioRel: towards large-scale biomedical relation extraction

BackgroundAlthough biomedical publications and literature are growing rapidly, there still lacks structured knowledge that can be easily processed by computer programs. In order to extract such knowledge from plain text and transform them into structural form, the relation extraction problem becomes an important issue. Datasets play a critical role in the development of relation extraction methods. However, existing relation extraction datasets in biomedical domain are mainly human-annotated, whose scales are usually limited due to their labor-intensive and time-consuming nature.ResultsWe construct BioRel, a large-scale dataset for biomedical relation extraction problem, by using Unified Medical Language System as knowledge base and Medline as corpus. We first identify mentions of entities in sentences of Medline and link them to Unified Medical Language System with Metamap. Then, we assign each sentence a relation label by using distant supervision. Finally, we adapt the state-of-the-art deep learning and statistical machine learning methods as baseline models and conduct comprehensive experiments on the BioRel dataset.ConclusionsBased on the extensive experimental results, we have shown that BioRel is a suitable large-scale datasets for biomedical relation extraction, which provides both reasonable baseline performance and many remaining challenges for both deep learning and statistical methods.

Influence of Quality of Remote Sensing Data on Vegetation Passability by Terrain Vehicles

The article studied databases of vegetation created from remote sensors, outcome of analyses of Cross-Country Movement in forests, and quality of utilized data. The aim was to combine various databases of forests and get statistics of best data by using different methods of evaluation. Passability in forests is mainly conducted with analysis of driving between trees. The most suitable datasets in the Czech Republic are Forest Economic Plan and Digital Elevation Model 5th generation combined with Digital Surface Model 1st generation. Accuracy and usability of databases were compared with digital model of surface created from orthophoto images. Processing of data is the most important part that influences quality of statistical and map results. Studied characteristics of input databases and applied methods also have considerable influence on results of analysis of forest passability. The outcome substantially varies for personnel armored vehicles and wheeled vehicles mostly due to their movement capabilities.

Diabetes Correlated Renal Fault Prediction through Deep Learning

INTRODUCTION: Diabetic nephropathy is one of the complications of diabetes that causes damage to kidneys. Deep learning techniques are widely used to predict different diseases.OBJECTIVES: The main aim of this work is to develop an effective prediction model using deep learning. To get an effective model, a suitable dataset is considered that comprises of features related to diabetic nephropathy.METHODS: Deep belief network (DBN) is the proposed deep learning technique which is compared with naive bayes, CART decision tree, logistic regression and support vector machine. DBN is composed of Restricted Boltzmann Machines (RBM). The algorithms are analysed based on evaluation measures like area under PR curve, area under ROC curve, gini coefficient and jaccard index.RESULTS: After comparison of all algorithms, it was observed that DBN has performed better in terms of AUROC, gini coefficient and jaccard index with values 0.8203, 0.6406 and 0.7777 respectively. But CART obtained better value of 0.9039 only for AUPR.CONCLUSION: The proposed technique has outperformed other techniques in terms of three metrics and is identified as the best performing algorithm. Hence, it is suggested to use DBN while predicting diabetic nephropathy.

Prevalence and prognosis in cancer therapy-related cardiovascular diseases from big data of electronic health record in Japan

Abstract Background Recent advance in cancer therapy has improved the cancer prognosis, which also increased the number of elder cancer survivors. Yet the anticancer agents have some serious side effects on cardiovascular system. Although there have been previous studies regarding cancer-specific and/or therapeutic agents-specific cardiovascular side effects in the selected patients, real-world data of the epidemiology for the impact of cardiovascular diseases (CVD) in cancer survivors is still limited due to the lack of comprehensive dataset. In this regard, the electronic health record (EHR), a big data, from the National Health Insurance in Japan provides a unique opportunity to obtain the suitable dataset. Purpose The purpose of this study was to clarify the prevalence of various CVD after the use of anticancer agents and the prognosis in cancer survivors using EHR in Japan. Methods This study was a retrospective cohort study using the database of Medical Data Vision (MDV) Co., Ltd. (Tokyo, Japan). The database covered 129 hospitals of all of the Designated Cancer Hospitals in Japan. The total number of patients was 17.85 million, which covered 14% of the total Japanese population. We extracted 140,327 cancer patients who received anticancer agents between April 2008 to January 2017. We identified CVD as pulmonary thromboembolism (PTE), atrial fibrillation (AF), ischemic heart disease (IHD), ventricular arrhythmia, pericarditis patients after the use of anticancer agents, according to International Classification of Diseases, 10th revision (ICD-10) codes. Hypertension (HT) patients were also identified according to ICD-10 codes, with the use of one or more anti-HT drugs after the HT diagnosis. Results The mean follow-up period was 1.53 years and mean age (standard deviation) was 69.2 (11.6) years old. The population over 75 years old were 36.2%, while males were 59.0%. There were prostate cancer (28.3%), lung cancer (13.8%), and colon cancer (13.2%) in males, and breast cancer (43.1%), colon cancer (12.6%), and lung cancer (8.6%) in females. Among them, 3,371 patients (2.4%) were diagnosed as PTE, 2,193 patients (1.6%) as AF, 4,085 patients (2.9%) as stable coronary artery disease, 681 patients (0.5%) as acute myocardial infarction (AMI), 749 patients (0.5%) as ventricular arrhythmia, 238 patients (0.2%) as pericarditis, 7,693 patients (5.5%) as HT. The cumulative survival rate was 65.6% in all cancer patients, 50.7% in PTE, 53.4% in AF, 61.7% in stable coronary disease, 52.7% in AMI and 45.9% in HT after the use of anticancer agents. Conclusion This study indicated that the prevalence of each CVD was less than 10% after the use of anticancer agents, but that patients with cancer therapy-related CVD had worse prognosis. Funding Acknowledgement Type of funding source: None

Dataset Splitting Techniques Comparison For Face Classification on CCTV Images

The performance of classification models in machine learning algorithms is influenced by many factors, one of which is dataset splitting method. To avoid overfitting, it is important to apply a suitable dataset splitting strategy. This study presents comparison of four dataset splitting techniques, namely Random Sub-sampling Validation (RSV), k-Fold Cross Validation (k-FCV), Bootstrap Validation (BV) and Moralis Lima Martin Validation (MLMV). This comparison is done in face classification on CCTV images using Convolutional Neural Network (CNN) algorithm and Support Vector Machine (SVM) algorithm. This study is also applied in two image datasets. The results of the comparison are reviewed by using model accuracy in training set, validation set and test set, also bias and variance of the model. The experiment shows that k-FCV technique has more stable performance and provide high accuracy on training set as well as good generalizations on validation set and test set. Meanwhile, data splitting using MLMV technique has lower performance than the other three techniques since it yields lower accuracy. This technique also shows higher bias and variance values and it builds overfitting models, especially when it is applied on validation set.

Dynamical study of quadrating harvesting of a predator–prey model with Monod–Haldane functional response

In this study, we have investigated local and global dynamics of a modified Leslie–Gower predator–prey model with Monod–Haldane functional response, where prey is subjected to quadratic harvesting. It is found that the solutions of the proposed system are positive and bounded uniformly. The feasible equilibrium points are also obtained for some suitable and predefined conditions. It is observed that the system exhibits at most three non-zero interior equilibrium points for different choices of parameters under certain conditions. The dynamics of all these feasible equilibrium points have been analysed using Routh–Hurwitz criterion. Local bifurcations analysis such as transcritical and saddle-node bifurcations have been investigated using Sotomayor’s theorem. To demonstrate the analytical results, numerical simulations using some suitable data set are carried out. Optimal harvesting policy has been obtained using Pontryagin’s Maximum Principle to show that the species can be preserved from extinction and a sustainable fishery can be achieved.

Fine-Grained Instance-Level Sketch-Based Image Retrieval

The problem of fine-grained sketch-based image retrieval (FG-SBIR) is defined and investigated in this paper. In FG-SBIR, free-hand human sketch images are used as queries to retrieve photo images containing the same object instances. It is thus a cross-domain (sketch to photo) instance-level retrieval task. It is an extremely challenging problem because (i) visual comparisons and matching need to be executed under large domain gap, i.e., from black and white line drawing sketches to colour photos; (ii) it requires to capture the fine-grained (dis)similarities of sketches and photo images while free-hand sketches drawn by different people present different levels of deformation and expressive interpretation; and (iii) annotated cross-domain fine-grained SBIR datasets are scarce, challenging many state-of-the-art machine learning techniques, particularly those based on deep learning. In this paper, for the first time, we address all these challenges, providing a step towards the capabilities that would underpin a commercial sketch-based object instance retrieval application. Specifically, a new large-scale FG-SBIR database is introduced which is carefully designed to reflect the real-world application scenarios. A deep cross-domain matching model is then formulated to solve the intrinsic drawing style variability, large domain gap issues, and capture instance-level discriminative features. It distinguishes itself by a carefully designed attention module. Extensive experiments on the new dataset demonstrate the effectiveness of the proposed model and validate the need for a rigorous definition of the FG-SBIR problem and collecting suitable datasets.

Fluid-Triggered Aftershocks in an Anisotropic Hydraulic Conductivity Geological Complex: The Case of the 2016 Amatrice Sequence, Italy

The mechanism by which faults interact each other is still a debated matter. One of the main issues is the role of pore-pressure diffusion in the delayed triggering of successive events. The 2016 Amatrice-Visso-Norcia seismic sequence (Central Apennines, Italy) provides a suitable dataset to test different physical mechanisms leading to delayed events. The sequence started on August 24, 2016, with the Amatrice mainshock (MW = 6), and was followed after more than 60 days by events in Visso (MW = 5.4) and Norcia (MW = 5.9). We analysed the contribution of the static stress change and the role of fluids in the delayed triggering. Through 3D poroelastic modelling, we show that the Amatrice mainshock induced a pore-pressure diffusion and a normal stress reduction in the hypocentral area of the two aftershocks, favouring the rupture. Our parametric study employs a simple two-layered conductivity model with anisotropy in the seismogenic layer, characterized by larger conductivity values (K>10-5 m/s) along the NNW-SSE direction. The one-way coupled pore-pressure 3-D diffusion modelling predicts the maximum increase of the pore pressure at the location of the two Visso earthquakes 60 days after the mainshock. The occurrence of anisotropic diffusivity is supported by the pattern of active faults and the strong crustal anisotropy documented by S-wave splitting analysis. We conclude that the temporal evolution of the sequence was controlled by the anisotropic diffusion of pore-pressure perturbations through pre-existing NNW-trending fracture systems.

Predicting Type 2 Diabetes A Machine Learning Approach

Diabetes is a well-known common disease among people around the world. Diabetes causes many anomalies in the body and results in the patients to become under a long term medication. Detecting diabetes has been done via hectic medical tests and causes a delay for the patients to get to know their test results. However, data mining and machine learning approaches are in the frontline supporting the health care domain to make effective predictions in this regard. This paper elaborates about predicting Type 2 Diabetes Mellitus using classification models. A suitable secondary dataset was used to build classification models and the more suitable model was selected via the valid performance measures. In this line, the Random Forest, Support Vector Machine, Naïve Bayes and Artificial Neural Network models were built. Based on the performance measures, Random Forest has been identified as the more suitable classifier with the accuracy of 90%, the recall and precision value of 0.90.

Application of a multi-stage neural network approach for time-series landfill gas modeling with missing data imputation

To mitigate the greenhouse gas effect, accurate and precise landfill gas prediction models are required for more precise prediction of the amount and recovery time of methane gas from landfills. When the study associates to greenhouse gas emissions problems, time series prediction models are of considerable interests, in which significant past records of gas data are required. This study is the first to specially impute the missing methane (CH4) data for applying in time series artificial neural network (ANN) model in an attempt to predict daily CH4 generation rate from a landfill in Regina, SK, Canada. Pre-processing was conducted on data to evaluate independent and significant meteorological input variables and provide suitable dataset for developing CH4 generation models. A two-stage time series model proposed in this study was performed by missing data imputation at the first stage, followed by a neural network auto-regressive model with exogenous inputs (NARX) at the second stage. The model with 3 layers, 5 climatic variables and 9 neurons in the hidden layer was the optimal structure. This model shows the high performance in CH4 prediction with the average index of agreement of 0.92 and the average mean absolute percentage error (MAPE) of 3.03% during the testing stage. Missing data imputation coupled with NARX method decreased the mean squared error (MSE) of the model by 84% (compared to Multilayer Perceptrons neural network model) in the testing period representing the effectiveness of missing data estimation coupling with time series ANN models in daily CH4 generation prediction.

Real-MFF: A large realistic multi-focus image dataset with ground truth

Multi-focus image fusion, a technique to generate an all-in-focus image from two or more partially-focused source images, can benefit many computer vision tasks. However, currently there is no large and realistic dataset to perform convincing evaluation and comparison of algorithms in multi-focus image fusion. Moreover, it is difficult to train a deep neural network for multi-focus image fusion without a suitable dataset. In this letter, we introduce a large and realistic multi-focus dataset called Real-MFF, which contains 710 pairs of source images with corresponding ground truth images. The dataset is generated by light field images, and both the source images and the ground truth images are realistic. To serve as both a well-established benchmark for existing multi-focus image fusion algorithms and an appropriate training dataset for future development of deep-learning-based methods, the dataset contains a variety of scenes, including buildings, plants, humans, shopping malls, squares and so on. We also evaluate 10 typical multi-focus algorithms on this dataset for the purpose of illustration.

TECHNIQUES FOR MOSAICS DOCUMENTATION THROUGH PHOTOGRAMMETRY DATA ACQUISITION. THE BYZANTINE MOSAICS OF THE NATIVITY CHURCH

Abstract. This paper describes a sequence of actions developed to guarantee a reliable and suitable dataset for the creation of detailed ortho images of Nativity Church mosaics. During acquisition campaigns, different photogrammetric techniques were tested, and different survey instruments were compared to improve the quality of the data obtained. The different outputs allow the adjustment of the instrument parameters and the acquisition methods, to structure a methodological process aimed at obtaining an accurate level of detail to describe the individual mosaic tile.From the realization of a reliable photomosaics an automatic vectorization system has been developed. This process, aimed at digitizing the tiles of the Church walls and the pavement mosaics, responds to a documentation and management purpose and to an objective of structuring a data acquisition method and post-production that can be replicated on other mosaic contexts.