Search

High-dimensional non-Gaussian data is widely found in the real world, such as in face recognition, facial expressions, document recognition, and text processing. Linear discriminant analysis (LDA) as dimensionality reduction performs poorly on non-Gaussian data and fails on high-dimensional data when the number of features is greater than the number of instances, commonly referred to as a small sample size (SSS) problem. We proposed a new method to reduce the number of dimensions called iterative LDA (iLDA). This method will handle the iterative use of LDA by gradually extracting features until the best separability is reached. The proposed method produces better vector projections than LDA for Gaussian and non-Gaussian data and avoids the singularity problem in high-dimensional data. Running LDA does not necessarily increase the excessive computational cost caused by calculating eigenvectors since the eigenvectors are calculated from small-dimensional matrices. The experimental results show performance improvement on 8 out of 10 small-dimensional datasets, and the best improvement occurs on the ULC dataset, from 0.753 to 0.861. For image datasets, accuracy improved in all datasets, with the Chest CT-Scan dataset showing the greatest improvement, followed by Georgia Tech from 0.6044 to 0.8384 and 0.8883 to 0.9481, respectively.

Materials informatics often struggles with small datasets. Our study introduces the Gaussian Mixture Model Virtual Sample Generation (GMM-VSG) approach to enhance feature correlation by generating virtual samples. Applied to six small and one large dataset of 218 N-heterocyclic compounds, GMM-VSG significantly improved predictive performance. Random Forest’s R2 rose from 0.80 to 0.99, with RMSE dropping from 9.87 to 0.22. Kernel Ridge’s R2 increased from 0.70 to 0.99, and RMSE decreased from 10.08 to 0.83. KNN improved from R2 of 0.74–0.90. ANN and MLPNN also saw notable improvements. GMM-VSG is thus crucial for advancing anti-corrosion material research.

In this investigation, a quantitative structure-property relationship (QSPR) model coupled with a quantum neural network (QNN) was used to explore the corrosion inhibition efficiency (CIE) of quinoxaline compounds. Integrating quantum chemical properties (QCP) features reduced computational burden by strategically reducing the features from 11 to 4 while maintaining prediction accuracy. QNN models outperform traditional methods like artificial neural networks (ANN) and multilayer perceptron neural networks (MLPNN), with a coefficient of determination (R2) value of 0.987, coupled with diminished root mean square error (RMSE), mean absolute error (MAE), and mean absolute deviation (MAD) values of 0.97, 0.92, and 1.10, respectively. Predictions for six newly synthesized quinoxaline derivatives: quinoxaline-6-carboxylic acid (Q1), methyl quinoxaline-6-carboxylate (Q2), (2E,3E)-2,3-dihydrazono-1,2,3,4-tetrahydroquinoxaline (Q3), (2E,3E) 2,3-dihydrazono-6-methyl-1,2,3,4-tetrahydroquinoxaline (Q4), (E)-3-(4-methoxyethyl)-7-methylquinoxalin-2(1 H)-one (Q5), and 2-(4-methoxyphenyl)-7-methylthieno[3,2-b] quinoxaline (Q6), show remarkable CIE values of 95.12, 96.72, 91.02, 92.43, 89.58, and 93.63 %, respectively. This breakthrough technique simplifies testing and production procedures for new anti-corrosion materials.

Text summarization is the process of creating a text summary that contains important information from a text document. In recent years, significant progress has been made in the field of text summarization research, along with the challenges that drive research progress in the field at large. The development of textual data has sparked great interest in text summarization research, which is thoroughly reviewed in this survey study. Text summarization research improvements continue to be made to date with various approaches, such as abstractive and extractive. The abstractive approach uses an intermediate representation of the input document to produce a summary that may differ from the original text. The extractive approach means that key sentences are extracted from the source document and combined to form a summary. Despite the various methodologies and approaches recommended, the summaries produced still contain ambiguities that can be interpreted with different meanings, resulting in errors in defining ambiguities, uncertainty in measuring the quality of summaries, difficulty in modeling linguistic context, difficulty in representing semantic meanings, and difficulty in specifying types of ambiguities. This research survey offers a comprehensive exploration of text summarization research, covering challenges, classifications, approaches, preprocessing methods, features, techniques, and evaluation methods, meeting future research needs. The results provide an overview of the state of the art of recent research developments in the topic of ambiguity resolution in text summarization, such as trends in research topics and approaches or techniques used in addressing ambiguity problems in text summarization.

The aspect-based sentiment analysis of hotel reviews represents a critical yet underexplored area of research. Such reviews encompass a range of expressions that extend beyond simple positive, negative, or neutral sentiments, delving into specific aspects discussed within the review. For instance, a statement like 'the location is comfortable, the food is good, but unfortunately, the service is bad' encapsulates multiple aspects, including location, food, and service. Existing research in Aspect-Based Sentiment Analysis (ABSA) frequently employs word embedding techniques for feature extraction, utilizing both Term and Transformer models. This study introduces an innovative approach by transitioning from word to sentence embedding, employing the SBERT transformer model for this transformation. This study further innovates label creation and topic modeling by integrating Bayesian search clustering with Inverse Document Frequency (IDF) calculations. This approach enables the identification of relevant topics, categorizing IDF scores based on occurrence frequencies. Scores ranging between 4–6 are marked as having low value but high informative potential, making them suitable candidates for label-specific datasets, particularly those without pre-existing labels. Additionally, this study incorporates the BigBird model to enhance this classification methodology. Applying this comprehensive framework to hotel review datasets, the result of the ABSA evaluation achieved an impressive accuracy score of 99%, indicating the efficacy of the proposed methodology in this nuanced field.

This scientific paper aims to investigate the best machine learning (ML) for predicting the corrosion inhibition efficiency (CIE) value of amino acid compounds. The study applied a quantitative structure–property relationship (QSPR) model based on an ML approach to predict the CIE values of three new amino acid compounds, namely L-asparagine (LA), L-isoleucine (LI), and L-proline (LP). The result is that the Gradient Boosting Regressor (GBR) model is proven to be the best predictive model based on the coefficient of determination (R2) and root mean square error (RMSE) metrics used. The study found that the three amino acid compounds LA, LI, and LP tested had high CIE values, ranging from 90.49% to 93.67%. These results are also relevant to the CIE values resulting from experimental studies and show a trend that is by the adsorption energy trend. This engineering breakthrough can be used to predict the corrosion inhibition properties of new compounds before experimental synthesis.

Biopreservation using lactic acid bacteria (LAB) is a promising technology to prevent the growth of pathogenic microorganisms in fresh and mildly processed food. The main aim of this study was to select LAB, originally isolated from ready-to-eat (RTE) seafood, for biopreservation of fresh salmon and processed salmon products. Ten LAB strains (five Carnobacterium and five Leuconostoc) were selected based on previously demonstrated bioprotective properties to investigate their antimicrobial mechanisms and temperature-dependent growth kinetics in a sterile salmon juice model system. Furthermore, five strains (three Carnobacterium and two Leuconostoc) were selected to test process-dependent growth kinetic parameters relevant to the secondary processing of salmon. Two strains (Carnobacterium maltaromaticum 35 and C. divergens 468) showed bacteriocin-like activity against Listeria innocua, while inhibitory effect of cell-free supernatants (CFS) was not observed against Escherichia coli. All selected strains were able to grow in sterile salmon juice at tested temperatures (4, 8, 12 and 16 °C), with specific growth rates (μ) ranging from 0.01 to 0.04/h at 4 °C and reaching a maximum population density of 8.4–9 log CFU/ml. All five strains tested for process-dependent growth kinetic parameters were able to grow in the range of 0.5–5% NaCl and 0.13–0.26% purified condensed smoke (VTABB and JJT01), with inter- and intraspecies variation in growth kinetics. According to the temperature-dependent growth kinetics and antimicrobial assay results, two strains, Leuconostoc mesenteroides 68 (Le.m.68) and C. divergens 468 (C d.468), were selected for in situ test to validate their ability to grow in vacuum-packed fresh salmon at 4 °C. Both strains were able to grow at maximum growth rates of 0.29 ± 0.04/d for Le. m.68 and 0.39 ± 0.06/d for C.d.468, and their final concentrations were 7.91 ± 0.31 and 8.02 ± 0.25 log CFU/g, respectively. This study shows that LAB, originally isolated from RTE seafood, have promising potential as bioprotective strains in fresh and processed salmon products.

Due to the continuous growth of the world population, there is an urgent need to find sustainable sources of high-quality protein. Fish side streams rich in essential nutrients and accounting for 60–70% of the whole fish, represent a sustainable source for recovery of valuable protein compounds. The present study aimed at extensive characterization of physicochemical, antioxidant and techno-functional properties of fish protein hydrolysate (FPH) obtained from farmed rainbow trout (Oncorhynchus mykiss). The FPH was produced from a minced rainbow trout raw material by enzymatic hydrolysis performed at 50 °C with addition of 0.05% w/w papain and 0.05% w/w bromelain. After inactivation of the proteases at 90 °C for 10 min, the content of the bioreactor was centrifuged, and the soluble protein fraction (FPH) was collected and freeze-dried. The total protein content of the FPH with 17.24% degree of hydrolysis was high (88.9%) and mainly represented by water-soluble proteins, while the lipid content was below 1%. In addition to the high protein content, trout hydrolysate had low protein oxidation values characterized by a relatively low total carbonyl content together with high amount of thiol groups (3.64 ± 0.31 and 20.7 ± 0.6 nmol/mg protein, respectively). No glass transition was detected in the differential scanning calorimetry (DSC) heat flow curves, suggesting lack of unfreezable solution formation in the FPH at freezing temperatures. The viscosity of FPH showed typical Newtonian behaviour. A peptidomic investigation (using HPLC-MS/MS technique) displayed chemical composition of the trout hydrolysate and identified peptide sequences which are present in the hydrolysate mixture, as well as proteins to which each peptide belongs to. In conclusion, it was suggested to use the obtained trout hydrolysate as a functional ingredient in the food and nutraceutical industry.

The main objective of this study was to investigate how the freshness of saithe (Pollachius virens) viscera affected the quality, composition and yield of oil obtained by silaging. Minced viscera with and without liver were stored separately for up to 3 days at 4 °C before silaging at pH 3.8 for 6 days at 10 °C. An antioxidant mixture was added to evaluate the effect on the lipid oxidation. Oil was extracted thermally from untreated raw material during storage (day 0–3) and after silaging. For oil obtained after silaging of viscera with liver, the oil yields increased significantly when the raw material was stored for more than one day prior to the treatment. Use of fresh raw material (collected at day 0) led to significantly lower oxidation compared to longer raw material storage. After one day of storage, the oxidation was less dependent on the freshness. Silaging with antioxidants resulted in significantly lower formation of oxidation products compared to acid without antioxidants and the most significant differences were observed after one day of storage. Contents of docosahexaenoic acid (DHA) and total omega-3 fatty acids decreased significantly when the raw material was stored for 1–3 days prior to silaging compared to fresh raw material. Results obtained by high resolution nuclear magnetic resonance (NMR) spectroscopy indicated that oxidation of esterified DHA might explain the DHA decrease. The free fatty acid content was highest when fresh raw material was used and was most likely affected by the formation of cholesteryl esters observed in NMR spectra after longer storage. The study shows that although the oil quality is reduced during silaging, processing shortly after catch and use of antioxidants can optimize the quality resulting in less oxidized oil richer in omega-3 fatty acids.

Various image steganography methods have been developed to improve the stego image quality, i.e. imperceptibility, capacity, and security. This research proposes an adaptive method that can select the most optimal pattern to minimize the error ratio due to message embedding. This adaptive pattern can optimize the performance of the inverted LSB substitution method, based on the two-bit + least-significant-bit (LSB) pattern in the container image. Before embedding, the message and container image bits are tested and the error ratio is calculated using the inverted LSB substitution method for various possible patterns. The pattern that has the least error rate is selected to embed the message. The use of this adaptive pattern in inverted LSB image steganography significantly increases imperceptibility. Based on the test results and by comparing it to previous studies, it is found increased results, the PSNR value ranges from 52.49 to 57.45, and the SSIM ranges from 0.9991 to 0.9999 with 1 BPP capacity.

Text summarization automatically produces a summary containing important sentences and includes all relevant important information from the original document. One of the main approaches, when viewed from the summary results, are extractive and abstractive. An extractive summary is heading towards maturity and now research has shifted towards abstractive summation and real-time summarization. Although there have been so many achievements in the acquisition of datasets, methods, and techniques published, there are not many papers that can provide a broad picture of the current state of research in this field. This paper provides a broad and systematic review of research in the field of text summarization published from 2008 to 2019. There are 85 journal and conference publications which are the results of the extraction of selected studies for identification and analysis to describe research topics/trends, datasets, preprocessing, features, techniques, methods, evaluations, and problems in this field of research. The results of the analysis provide an in-depth explanation of the topics/trends that are the focus of their research in the field of text summarization; provide references to public datasets, preprocessing and features that have been used; describes the techniques and methods that are often used by researchers as a comparison and means for developing methods. At the end of this paper, several recommendations for opportunities and challenges related to text summarization research are mentioned.

Shopping centres function as meeting places, and people are using them for various non-shopping purposes. Knowledge about different social spaces, and how they are used and perceived, is important for understanding health and wellbeing in the community and for developing health promoting societies. This scoping review provides an overview of existing research examining people's positive experiences and motivations for visiting shopping centres when shopping is not the main purpose. The results showed that people are motivated by the variety of available activities found in shopping centres and that they valued specific site features of the shopping centres. People reported a range of positive experiences when visiting shopping centres including social interactions, opportunities to escape their everyday life, and experiences of place attachment and social cohesion. This shows that, non-shopping use of shopping centres seems to be important for people's wellbeing and might facilitate health promotion and social sustainability in local communities.

Summary: The emergence of SARS-CoV-2 variants of concern (VOC) is driven by mutations that mediate escape from neutralizing antibodies. There is also evidence that mutations can cause loss of T cell epitopes. However, studies on viral escape from T cell immunity have been hampered by uncertain estimates of epitope prevalence. Here, we map and quantify CD8 T cell responses to SARS-CoV-2-specific minimal epitopes in blood drawn from April to June 2020 from 83 COVID-19 convalescents. Among 37 HLA ligands eluted from five prevalent alleles and an additional 86 predicted binders, we identify 29 epitopes with an immunoprevalence ranging from 3% to 100% among individuals expressing the relevant HLA allele. Mutations in VOC are reported in 10.3% of the epitopes, while 20.6% of the non-immunogenic peptides are mutated in VOC. The nine most prevalent epitopes are conserved in VOC. Thus, comprehensive mapping of epitope prevalence does not provide evidence that mutations in VOC are driven by escape of T cell immunity.

Collaboration is an important 21st century skill, and there is an increasing interest in computer-supported collaborative learning (CSCL) in teaching and learning practices regarding the necessary knowledge and skills for 21st century learners. However, solutions are still lacking concerning how to enable learners to collaborate when working with a specific subject. This study developed and tested a learning design using Minecraft as an educational tool in math classes in higher education over a 2-year period. The research questions are: 1) How can Minecraft be used as an educational tool for learning subject-specific skills, and what are the implications for the development of 21st century skills? 2) How can social network analysis (SNA) and interaction analysis (IA) be used as a research methodology for analyzing the use of Minecraft as an educational tool for learning subject-specific skills? Data was analyzed using a mixed-methods approach, in which both a quantitative perspective through SNA and a qualitative perspective through IA were employed. Findings of the study: 1) suggested that using Minecraft for learning a subject-specific skill triggered collaborative learning processes, 2) facilitated the development of 21st century skills in math classes, and 3) combining SNA and IA as a research methodology for analyzing the use of Minecraft for learning subject-specific skills. These findings serve as a stepping stone for teachers and researchers to create learning designs that use a digital educational tool to facilitate collaborative learning and advance the literature on CSCL and educational practice.

The effect of freshness of heads of rainbow trout (Onchorhynchus mykiss) for hydrolysing fish protein was investigated. To simulate storage and transportation, hydrogen peroxide and iron was added to minced heads to induce oxidation in the raw material prior to one week of storage. The effect of antioxidant in retarding oxidative changes during hydrolysis or to the raw material was investigated by adding butylated hydroxy toluene (BHT) prior to hydrolysis or storage. Enzymatic hydrolysis was carried out using bromelain and papain. The oil fraction was separated from the water soluble proteins, and the soluble phase was freeze dried. Both the oil fraction and protein fraction from enzymatic hydrolysis was affected by oxidative state of raw material. FFA was significantly higher in those FPH made from raw material added pro-oxidants, addition of antioxidant did not affect the level of FFA. The solubility of proteins in dried fish protein hydrolysates (FPH) decreased significantly when using oxidized raw material. Although addition of antioxidant improved the solubility, it was still significantly lower compared to those FPHs not added pro-oxidants. The FPH with decreased solubility also had higher levels of carbonyl groups which indicate protein oxidation. However, the oxidative state of raw material did not affect fatty acid composition in oil fraction or the amino acid composition in the FPH.

A combined effect of pulsed electric field application and salting in a brine with 5 and 10% w/w NaCl on oxidative stability of lipids and proteins, as well as color characteristics of sea bass samples, was assessed in the study. The applied intensity of the current was set at 10 and 20 A corresponding to 300 and 600 V cm−1, respectively. Pulsed electric field (PEF) treatment led to a significant (p < 0.05) increase in primary and secondary lipid oxidation products expressed as peroxide value, conjugated dienes and 2-thiobarbituric acid reactive substances in PEF-treated samples compared to untreated ones. Conjugated dienes, as unstable primary oxidation products, correlated with b∗-value (p < 0.05, R = 0.789), suggesting their contribution to the yellowness of the fish flesh due to fast decomposition and conversion into secondary oxidation products yielding yellow pigmentation.However, none of the fish samples treated at the higher current intensity of 20 A exceeded the acceptable level of 5 meq active oxygen/kg lipid according to the requirements of the Standard for fish oils CODEX STAN 329–2017, suggesting acceptable oxidative status quality of sea bass samples after the treatment. PEF-treated fish samples also showed a significant increase in Schiff bases and total carbonyls on day 5 and day 8 of brine salting compared to non-treated samples, revealing a strong effect of electroporation on protein oxidation.

The data article refers to the paper “Semi-dry storage as a maturation process for improving the sensory characteristics of the edible red seaweed dulse (Palmaria palmata)” [1]. The data refers to the analysis of samples of the edible seaweed species Palmaria palmata during storage in a dry (D, containing ca. 6 % moisture) and semi-dry state (SD, containing ca. 20 % moisture). The article includes data from the analysis of samples taken at 0, 12, 61 and 126 days of storage to evaluate the effect of moisture content and storage time on the sensory characteristics of the product. The variations in flavor, odor and texture between samples were measured by sensory evaluation. Data from the analysis of flavor-active compounds (free amino acids and volatile compounds), macronutrient content (soluble proteins and carbohydrates, lipid and mineral fractions), physico-chemical properties (water activity, water and oil-binding capacities, swelling capacity), color and microbial load are also reported. The information provided in this article can be used by industrial stakeholders (seaweed producers, food industry) to optimize processing and storage conditions of edible seaweeds and by scientists to build upon further knowledge to improve the quality of seaweeds in food applications.

The neuronal ceroid lipofuscinoses (NCLs) are devastating inherited progressive neurodegenerative diseases, with most forms having a childhood onset of clinical signs. The NCLs are characterized by progressive cognitive and motor decline, vision loss, seizures, respiratory and swallowing impairment, and ultimately premature death. Different forms of NCL result from mutations in at least 13 genes. The clinical signs of some forms overlap significantly, so genetic testing is the only way to definitively determine which form an individual patient suffers from. At present, an effective treatment is available for only one form of NCL. Evidence of NCL has been documented in over 20 canine breeds and in mixed-breed dogs. To date, 12 mutations in 8 different genes orthologous to the human NCL genes have been found to underlie NCL in a variety of dog breeds. A Dachshund model with a null mutation in one of these genes is being utilized to investigate potential therapeutic interventions, including enzyme replacement and gene therapies. Demonstration of the efficacy of enzyme replacement therapy in this model led to successful completion of human clinical trials of this treatment. Further research into the other canine NCLs, with in-depth characterization and understanding of the disease processes, will likely lead to the development of successful therapeutic interventions for additional forms of NCL, for both human patients and animals with these disorders.

Bioactivities of bulk fish protein hydrolysates (FPH) from defatted salmon backbones obtained with eight different commercial enzymes and their combinations were tested. All FPH showed antioxidative activity in vitro. DPPH scavenging activity increased, while iron chelating ability decreased with increasing time of hydrolysis. All FPH showed ACE inhibiting effect which depended on type of enzyme and increased with time of hydrolysis. The highest effect was found for FPH produced with Trypsin. Bromelain + Papain hydrolysates reduced the uptake of radiolabelled glucose into CaCo-2 cells, a model of human enterocytes, indicating a potential antidiabetic effect of FPH. FPH obtained by Trypsin, Bromelain + Papain and Protamex showed the highest ACE inhibitory, cellular glucose transporter (GLUT/SGLT) inhibitory and in vitro antioxidative activities, respectively. Correlation was observed between the measured bioactivities, degree of hydrolysis and molecular weight profiles, supporting prolonged hydrolysis to obtain high bioactivities.

The synthesis of a naphthylvinyl-labeled glycerol ether analog of phosphatidylcholine, 1-O-[12-(2-naphthyl)-dodec-11-enyl]-2-O-decanoyl-sn-glycerol-3- phosphocholine (NVPC), is described. This involves a Wittig reaction between 2-naphthaldehyde and a phosphonium salt which gives the trans-naphthylvinyl group as the predominant isomer. Lyso NVPC was prepared from NVPC by phospholipase A2 action. NVPC absorbs strongly at 248 nm (epsilon = 58,300 M-1 cm-1) and gives broad fluorescence emission with maxima at 343 nm and 360 nm and a quantum yield of 0.10 in ethanol. An assay for phospholipase A2 was developed using high performance liquid chromatography with fluorescence detection to separate and quantify NVPC and lyso NVPC. Activities as low as 1-2 pmol/min in an assay volume of 0.1 ml can easily be measured. The assay was used with a pure enzyme from cobra venom and a crude enzyme from synovial fluid. Enzyme specificities for phosphatidylcholine and NVPC with cobra venom and porcine pancreatic phospholipases A2 were compared using a titrametric assay. The use of the assay with NVPC to study the metabolism of platelet activating factor is discussed.