In this paper, we propose using Learning from Answer Sets to approximate black-box models, such as Neural Networks (NN), in the specific case of learning user preferences. We specifically explore the use of ILASP (Inductive Learning of Answer Set Programs) to approximate preference learning systems through weak constraints. We have created a dataset on user preferences over a set of recipes, which is used to train the NNs that we aim to approximate with ILASP. Our experiments investigate ILASP both as a global and a local approximator of the NNs. These experiments address the challenge of approximating NNs working on increasingly high-dimensional feature spaces while achieving appropriate fidelity on the target model and limiting the increase in computational time. To handle this challenge, we propose a preprocessing step that exploits Principal Component Analysis to reduce the dataset’s dimensionality while keeping our explanations transparent.

Under consideration for publication in Theory and Practice of Logic Programming (TPLP).

This study examines the potential influence of deformation on the systematics of Rb-Sr geochronology in mica phases under different conditions. Biotite and muscovite porphyroclasts in deformed specimens were characterized using electron backscattered diffraction, electron probe microanalysis and laser ablation inductively coupled plasma mass spectrometry to quantify spatial variations in crystal lattice orientations, element concentrations and in situ Rb-Sr geochronology. S29, a specimen subjected to deformation at greenschist facies conditions, is characterized by a spread in in situ Rb-Sr two-point isochron spot dates, which exhibit a strong inverse correlation with lattice deformation. As such, these Rb-Sr dates are interpreted to record partial re-equilibration controlled by deformation. Rb-Sr data from white mica in a specimen (NP17-58), which was deformed at lower amphibolite facies conditions, define a single population isochron. No correlation between lattice distortion and Rb-Sr spot dates is noted. Finally, two biotite porphyroclasts and matrix grains in a specimen (AC4), deformed at upper amphibolite facies conditions, define unique, single population Rb-Sr isochrons. The Rb-Sr systematics of the older porphyroclast are interpreted to be mainly temperature-controlled. In contrast, the Rb-Sr systematics for the younger porphyroclast and matrix grains are interpreted to reflect fluid-mediated resetting. The results of this study demonstrate that the multi-faceted influences on Rb-Sr systematics make isolating the effect of deformation difficult. Due to the complexity of the Rb-Sr systematics in deformed specimens, careful consideration of the mica phase analysed, as well as the temperatures, fluids and deformation experienced throughout the rock’s history, needs to be accounted for.

While chrono-nutrition behaviours are inter-related, few studies examined patterns of chrono-nutrition behaviours using a comprehensive set of these behaviours. This study aimed to identify chrono-nutrition behaviour patterns and examine their associations with sociodemographic characteristics, diet quality and obesity. This cross-sectional study included 1047 Japanese adults aged 20–69 years. Using 11-d diaries of eating, chrono-nutrition behaviours (such as frequency and timing of eating) were evaluated for workdays and non-workdays separately. Principal component analysis identified four patterns: ‘early, large breakfast on workdays’, ‘skipping breakfast on non-workdays’, ‘frequent snacking with small dinner’ and ‘early last eating with large lunch’. Female sex was associated with the ‘frequent snacking with small dinner’ and ‘early last eating with large lunch’ patterns; male sex was associated with the ‘skipping breakfast on non-workdays’ pattern. Age was positively associated with the ‘skipping breakfast on non-workdays’ and ‘early last eating with large lunch’ patterns. Having a full-time paid job was associated positively with the two patterns characterised mainly by breakfast but inversely with the remaining two patterns. After adjustment for potential confounders, none of the four patterns were significantly associated with diet quality (Healthy Eating Index-2020 score), general obesity (BMI ≥ 25 kg/m2) or abdominal obesity (waist circumference ≥ 90 cm for males; ≥ 80 cm for females). In conclusion, this study suggests that different chrono-nutrition behaviour patterns were differentially associated with sociodemographic characteristics, but not with diet quality or obesity. Further research is needed to clarify if the patterning approach is useful to comprehensively interrogate chrono-nutrition behaviours.

This article investigates phonological complexity by using artificial neural network and Bayesian structural equation models to derive representations of phonological complexity from counts of the segments associated with particular features in languages’ phonemic inventories. These latent representations can then be used alongside principal component analysis to further analyse how interactions between phonological features affect overall complexity, and what phonological complexity patterns a model can detect in a phonological feature data set. The results indicate that the per-feature segment counts investigated tend to contribute positively to a language’s complexity, and that the latent complexity variables approximate a log-normal distribution. This implies that phonological complexifications co-occur with other complexifications diachronically while tending to be more constrained at the upper and lower ends of the complexity range.

This study aimed to evaluate the influence of different intermittent fasting regimens on metabolic parameters in healthy rats and compare them with caloric restriction. A total of fifty adult male Wistar rats (±90 days old) were randomised into 5 groups: control group (CON), caloric restriction group (CR), time-restricted feeding group (TRF), alternate-day fasting (ADF) group and alternate-day modified fasting group (ADMF). ADF and ADMF stood out for improving the metabolic parameters in healthy rats by presenting improvements in glucose parameters, greatest weight loss (ADF v. CON: −16·50 (sd 6·16) g; effect size = −5·34; 95 % CI: −7·05, −3·04; P < 0·001; ADMF v. CON: –21·88 (sd 6·66) g; effect size = −5·83; 95 % CI: −7·66, −3·36; P < 0·001) and higher HDL (ADF v. CON: 141·50 (sd 10·17) mg/dl; effect size = 3·03; 95 % CI: 1·01, 4·45; P < 0·001; ADMF v. CON: 133·10 (sd 5·94) mg/dl; effect size = 3·37; 95 % CI: 1·22, 4·86; P = 0·004). Additionally, ADMF presented a smaller adipocyte area among the fasting regimens (13·92 (sd 2·06) area/µm2; effect size = −4·20; 95 % CI: −5·45, −2·66; P < 0·001 v. CON), in addition to presenting muscle fibre hypertrophy (71·20 (sd 5·16) area/µm2; effect size = 2·93; 95 % CI: 1·57, 4·05; P < 0·001 v. CON), followed by ADF (adipocyte area: 19·25 (sd 0·87) area/µm2; effect size = −2·19; 95 % CI: −3·12, −1·12; P = 0·003 v. CON; muscle fibre: 53·80 (sd 6·61) area/µm2; effect size = 2·93; 95 % CI: 1·57, 4·05; P = 0·566 v. ADMF). The ADF and ADMF groups were more effective among the intermittent fasting regimens analysed in promoting improvements in metabolic parameters in healthy rats.

Good keeping quality (KQ) is a critical trait for sustaining potato cultivation under subtropical conditions, where post-harvest losses significantly impact profitability. To support breeding for improved KQ, a targeted evaluation of variability in key contributing traits was undertaken using a diverse germplasm set of 540 accessions of Solanum tuberosum ssp. tuberosum. The study utilized data from 2010 to 2020, incorporating control varieties Kufri Pukhraj, Kufri Dewa and Kufri Ashoka. Evaluation was performed using an augmented design with appropriate data transformations to mitigate annual environmental variations treated as block effects. The adjusted means revealed substantial phenotypic variation in sprouting (34.93%), firmness (20.77%), weight loss (27.32%), rottage (75.43%) and total weight loss (25.44%). Significant genotypic differences were observed for total weight loss and sprouting. Principal component analysis reduced data dimensionality, with the first three components accounting for 86.3% of the total variance. Biplots were generated using eigenvalues and eigenvectors to visualize the distribution of accessions based on KQ traits. Genotypes clustered in favourable zones on the biplots, enabling the identification of 18 superior keeping germplasm accessions: CP3151, CP3134, CP3117, CP3208, CP3211, CP3590, CP3515, CP3702, CP3336, CP3661, CP3514, CP4214, CP4229, CP4514, CP3588, CP3639, CP3795 and Kufri Dewa. The findings identify valuable parental material for breeding programs targeting improved post-harvest resilience in potato cultivars suited to the subtropical plains.

The Southern Ocean, a region characterized by high nutrient levels but often low productivity, hosts dynamic picophytoplankton communities crucial for its food web. This study investigated the spatial and inter-annual variability of picophytoplankton abundances and their environmental drivers in the Indian sector of the Southern Ocean during the austral summers of 2018 and 2020. Using flow cytometry for picophytoplankton quantification and standard oceanographic methods for environmental parameters (temperature, salinity, nitrate, phosphate, silicate), we employed descriptive statistics, inferential group comparisons (t-tests, analysis of variance), principal component analysis (PCA) and principal component regression (PCR) to analyse the dataset. Our analyses revealed significant differences in picophytoplankton abundances and environmental conditions across distinct oceanic fronts, between deep chlorophyll maximum and surface depths and, notably, between the two study years. PCA identified three major environmental gradients explaining over 93.5% of the variance in temperature, salinity, nitrate, phosphate and silicate. PCR confirmed our hypothesis: the abundance and carbon biomass of picoeukaryote II (PEUK-II) picophytoplankton was statistically significant overall (F-statistic = 3.415, P = 0.0290). The model explained 24.2% of the variance in PEUK-II abundance (R2 = 0.242), indicating its sensitivity to dynamic oceanographic conditions, with PC3 (primarily representing a salinity gradient) being a significant predictor. Conversely, Prochlorococcus-like/Synechococcus picophytoplankton abundance was not statistically significant overall (F-statistic = 2.068, P = 0.124), suggesting control by other, potentially non-linear factors. These findings highlight distinct ecological strategies among picophytoplankton groups and are vital for predicting their roles in the Southern Ocean’s microbial food web amidst ongoing environmental change.

Poor socket fit is the leading cause of prosthetic limb discomfort. However, currently clinicians have limited objective data to support and improve socket design. Finite element analysis predictions might help improve the fit, but this requires internal and external anatomy models. While external 3D surface scans are often collected in routine clinical computer-aided design practice, detailed internal anatomy imaging (e.g., MRI or CT) is not. We present a prototype statistical shape model (SSM) describing the transtibial amputated residual limb, generated using a sparse dataset of 33 MRI and CT scans. To describe the maximal shape variance, training scans are size-normalized to their estimated intact tibia length. A mean limb is calculated and principal component analysis used to extract the principal modes of shape variation. In an illustrative use case, the model is interrogated to predict internal bone shapes given a skin surface shape. The model attributes ~52% of shape variance to amputation height and ~17% to slender-bulbous soft tissue profile. In cross-validation, left-out shapes influenced the mean by 0.14–0.88 mm root mean square error (RMSE) surface deviation (median 0.42 mm), and left-out shapes were recreated with 1.82–5.75 mm RMSE (median 3.40 mm). Linear regression between mode scores from skin-only- and full-model SSMs allowed prediction of bone shapes from the skin with 3.56–10.9 mm RMSE (median 6.66 mm). The model showed the feasibility of predicting bone shapes from surface scans, which addresses a key barrier to implementing simulation within clinical practice, and enables more representative prosthetic biomechanics research.

Microwaves (MWs) have emerged as a promising sensing technology to complement optical methods for monitoring floating plastic litter. This study uses machine learning (ML) to identify optimal MW frequencies for detecting floating macroplastics (>5 cm) across S, C, and X-bands. Data were obtained from dedicated wideband backscattering radio measurements conducted in a controlled indoor scenario that mimics deep-sea conditions. The paper presents new strategies to directly analyze the frequency domain signals using ML algorithms, instead of generating an image from those signals and analyzing the image. We propose two ML workflows, one unsupervised, to characterize the difference in feature importance across the measured MW spectrum, and the other supervised, based on multilayer perceptron, to study the detection accuracy in unseen data. For the tested conditions, the backscatter response of the plastic litter is optimal at X-band frequencies, achieving accuracies up to 90% and 80% for lower and higher water wave heights, respectively. Multiclass classification is also investigated to distinguish between different types of plastic targets. ML results are interpreted in terms of the physical phenomena obtained through numerical analysis, and quantified through an energy-based metric.

Changes in waxed dry cheese during the ripening process, over periods of 7 and 30 days, were analysed using near-infrared spectroscopy (FT-NIR) and mid-infrared spectroscopy (FT-MIR) by attenuated total reflection (ATR). FT-NIR was employed to determine the proximate composition of the cheese (protein, fat, moisture, total solids, and salt content), identifying changes directly associated with the ripening process. FT-MIR data were used to identify spectral bands associated with chemical changes occurring during the cheese maturation. Additionally, chemometric techniques were applied to demonstrate the potential of FT-MIR infrared spectroscopy for cheese differentiation and fingerprint profiling. Subsequently, partial least squares discriminant analysis (PLS-DA) of the FT-MIR spectra was performed, revealing two distinct clusters representing the cheese ripening times. Functional groups related to lipids (–CH2 – and – CH3), proteins (amide bands I and II), and carbohydrates (C–O) were identified, correlating to lipolysis, proteolysis, and lactose catabolism. Infrared spectroscopy in combination with chemometric methods proved to be a robust and reliable tool for monitoring changes during the ripening of waxed dry cheese. The results obtained highlight its usefulness as an alternative approach for the analysis and fingerprinting of traditional Mexican foods, aiming to add value to local products.

Avocado is a delicious fruit crop having great economic importance. Understanding the extent of variability present in the existing germplasm is important to identify genotypes with specific traits and their utilization in crop improvement. The information on genetic variability with respect to morphological and biochemical traits in Indian avocados is limited and as it has hindered genetic improvement of the crop. In the current study, 83 avocado accessions from different regions of India were assessed for important 17 morphological and 8 biochemical traits. The results showed the existence of wide variability for traits such as fruit weight (75.88–934.12 g), pulp weight (48.08–736.19 g), seed weight (6.37–32.62 g), FRAP activity (27.65–119.81 mg AEAC/100 g), total carotenoids (0.96–7.17 mg/100 g), oil content (4.91–25.49%) and crude fibre (6.85–20.75%) in the studied accessions. The first three components of principal component analysis explained 54.79 per cent of total variance. Traits such as fruit weight, pulp weight, seed weight, moisture and oil content contributed more significantly towards total variance compared to other traits. The dendrogram constructed based on Euclidean distance wards minimum variance method divided 83 accessions into two major groups and nine sub clusters suggesting wide variability in the accessions with respect to studied traits. In this study, superior accessions for important traits such as fruit size (PA-102, PA-012), high pulp recovery (PA-036, PA-082,), thick peel (PA-084, PA-043, PA-011, PA-008), high carotenoids (PA-026, PA-096) and high oil content (PA-044, PA-043, PA-046, PA-045) were identified which have potential utility in further crop improvement programmes.

Perilla is a self-fertilizing crop widely used in East Asia for its seeds and leaves. Of the two varieties of Perilla, P. frutescens var. frutescens has long been used as a folk plant in South Korea. The seeds are rich in unsaturated fatty acids, which offer significant health benefits, making them popular for use in seed oil or as a spice. The leaves, with their high perilla ketone content and unique aroma, are used as leafy vegetables and spices. The morphological characteristics of crops are complex for various reasons, such as environment factors, multiplicity, etc. To better understand the morphological variations among three types of Perilla collected from three regions of South Korea, 7 qualitative traits and 10 quantitative traits were investigated using 500 Perilla accessions. The results of principal component analysis (PCA) indicated that the first two components together explained 52.2% of the overall variation. The 500 Perilla accessions clearly distinguished cultivated var. frutescens from weedy var. crispa and also revealed differences between cultivated and weedy types of var. frutescens. Significant morphological differences were observed among the three types of Perilla, especially in seed and plant characteristics. When the PCA results were analysed by region, regional differences were observed for all three types of Perilla. Therefore, this study provides a better understanding of the morphological and geographical differences in Perilla grown and naturally occurring in South Korea, which will aid research on crop evolution and differentiation, as well as Perilla breeding programmes.

Recent research into vowel covariation has suggested that speakers can be identified as leaders or laggers in multiple ongoing sound changes. What remains unclear is how stable a speaker’s patterns of covariation are over time and whether these leaders and laggers of sound changes remain leaders and laggers over time. We employ corpus data from 51 New Zealand English (NZE) speakers who were recorded at two time-points (eight years apart) and explore covariation between 10 monophthongs using principal component analysis (PCA). The results indicate significant stability across the time-points in two unique vowel clusters, suggesting that speakers’ covariation position within their community remains stable over time. The overall covariation patterns also replicate patterns previously observed in a different corpus of NZE, indicating that patterns of vowel covariation observed with PCA can be stable and replicable across multiple corpora.

Assessing children’s diets is currently challenging and burdensome. Abbreviated FFQ have the potential to assess dietary patterns in a rapid and standardised manner. Using nationally representative UK dietary intake and biomarker data, we developed an abbreviated FFQ to calculate dietary quality scores for pre-school and primary school-aged children. UK National Diet and Nutrition Survey (2008–2016) weekly consumption frequencies of 129 food groups from 4-d diaries were cross-sectionally analysed using principal component analysis. A 129-item score was derived, alongside a 12-item score based on foods with the six highest and six lowest coefficients. Participants included 1069 pre-schoolers and 2565 primary schoolchildren. The first principal component explained 3·4 and 3·0 % of the variation in the original diet variables for pre-school and primary school groups, respectively, and described a prudent diet pattern. Prudent diet scores were characterised by greater consumption of fruit, vegetables and tap water and lower consumption of crisps, manufactured coated chicken/turkey products, purchased chips and soft drinks for both age groups. Correlations between the 129-item and 12-item scores were 0·86 and 0·84 for pre-school and primary school-aged children, respectively. Bland–Altman mean differences between the scores were 0·00 sd; 95 % limits of agreement were −1·05 to 1·05 and −1·10 to 1·10 sd for pre-school and primary school-aged children, respectively. Correlations between dietary scores and nutritional biomarkers showed only minor attenuation for the 12-item compared with the 129-item scores, illustrating acceptable congruence between prudent diet scores. The two 12-item FFQ offer user-friendly tools to measure dietary quality among UK children.