To send content items to your account,
please confirm that you agree to abide by our usage policies.
If this is the first time you use this feature, you will be asked to authorise Cambridge Core to connect with your account.
Find out more about sending content to .
To send content items to your Kindle, first ensure email@example.com
is added to your Approved Personal Document E-mail List under your Personal Document Settings
on the Manage Your Content and Devices page of your Amazon account. Then enter the ‘name’ part
of your Kindle email address below.
Find out more about sending to your Kindle.
Note you can select to send to either the @free.kindle.com or @kindle.com variations.
‘@free.kindle.com’ emails are free but can only be sent to your device when it is connected to wi-fi.
‘@kindle.com’ emails can be delivered even when you are not connected to wi-fi, but note that service fees apply.
There is still little knowledge of objective suicide risk stratification.
This study aims to develop models using machine-learning approaches to predict suicide attempt (1) among survey participants in a nationally representative sample and (2) among participants with lifetime major depressive episodes. We used a cohort called the National Epidemiologic Survey on Alcohol and Related Conditions (NESARC) that was conducted in two waves and included a nationally representative sample of the adult population in the United States. Wave 1 involved 43 093 respondents and wave 2 involved 34 653 completed face-to-face reinterviews with wave 1 participants. Predictor variables included clinical, stressful life events, and sociodemographic variables from wave 1; outcome included suicide attempt between wave 1 and wave 2.
The model built with elastic net regularization distinguished individuals who had attempted suicide from those who had not with an area under the ROC curve (AUC) of 0.89, balanced accuracy 81.86%, specificity 89.22%, and sensitivity 74.51% for the general population. For participants with lifetime major depressive episodes, AUC was 0.89, balanced accuracy 81.64%, specificity 85.86%, and sensitivity 77.42%. The most important predictor variables were a diagnosis of borderline personality disorder, post-traumatic stress disorder, and being of Asian descent for the model in all participants; and previous suicide attempt, borderline personality disorder, and overnight stay in hospital because of depressive symptoms for the model in participants with lifetime major depressive episodes. Random forest and artificial neural networks had similar performance.
Risk for suicide attempt can be estimated with high accuracy.
Over the past two decades, early detection and early intervention in psychosis have become essential goals of psychiatry. However, clinical impressions are insufficient for predicting psychosis outcomes in clinical high-risk (CHR) individuals; a more rigorous and objective model is needed. This study aims to develop and internally validate a model for predicting the transition to psychosis within 10 years.
Two hundred and eight help-seeking individuals who fulfilled the CHR criteria were enrolled from the prospective, naturalistic cohort program for CHR at the Seoul Youth Clinic (SYC). The least absolute shrinkage and selection operator (LASSO)-penalized Cox regression was used to develop a predictive model for a psychotic transition. We performed k-means clustering and survival analysis to stratify the risk of psychosis.
The predictive model, which includes clinical and cognitive variables, identified the following six baseline variables as important predictors: 1-year percentage decrease in the Global Assessment of Functioning score, IQ, California Verbal Learning Test score, Strange Stories test score, and scores in two domains of the Social Functioning Scale. The predictive model showed a cross-validated Harrell's C-index of 0.78 and identified three subclusters with significantly different risk levels.
Overall, our predictive model showed a predictive ability and could facilitate a personalized therapeutic approach to different risks in high-risk individuals.
The coronavirus disease pandemic was initiated in Wuhan province of mainland China in December 2019 and has spread over the world.
This study analyses the effects of COVID 19 based on Likely Positive Cases and fatality in India during and after the lockdown period from 24 March 2020 to 24 May 2020.
Python has been used as the main programming language for data analysis and forecasting using the Prophet Model, a time series analysis model. The dataset has been preprocessed by grouping together the days for total numbers of cases and deaths on few selected dates and removed missing values present in some states.
The Prophet model performs better in terms of precision on the real data. Prediction depicts that during the lockdown, the total cases were rising but in a controlled manner with an accuracy of 87%. After the relaxation of lockdown rules, the predictions have shown an obstreperous situation with an accuracy of 60%.
The resilience could have been better if the lockdown with strict norms was continued without much relaxation. The situation after lockdown has been found to be uncertain as observed by the experimental study conducted in this work.
As the COVID-19 pandemic continues to escalate and place pressure on hospital system resources, a proper screening and risk stratification score is essential. We aimed to develop a risk score to identify patients with increased risk of COVID-19, allowing proper identification and allocation of limited resources. A retrospective study was conducted of 338 patients who were admitted to the hospital from the emergency room to regular floors and tested for COVID-19 at an acute care hospital in the Metropolitan Washington D.C. area. The dataset was split into development and validation sets with a ratio of 6:4. Demographics, presenting symptoms, sick contact, triage vital signs, initial laboratory and chest X-ray results were analysed to develop a prediction model for COVID-19 diagnosis. Multivariable logistic regression was performed in a stepwise fashion to develop a prediction model, and a scoring system was created based on the coefficients of the final model. Among 338 patients admitted to the hospital from the emergency room, 136 (40.2%) patients tested positive for COVID-19 and 202 (59.8%) patients tested negative. Sick contact with suspected or confirmed COVID-19 case (3 points), nursing facility residence (3 points), constitutional symptom (1 point), respiratory symptom (1 point), gastrointestinal symptom (1 point), obesity (1 point), hypoxia at triage (1 point) and leucocytosis (−1 point) were included in the prediction score. A risk score for COVID-19 diagnosis achieved area under the receiver operating characteristic curve of 0.87 (95% confidence interval (CI) 0.82–0.92) in the development dataset and 0.85 (95% CI 0.78–0.92) in the validation dataset. A risk prediction score for COVID-19 can be used as a supplemental tool to assist clinical decision to triage, test and quarantine patients admitted to the hospital from the emergency room.
Depression is a challenge to diagnose reliably and the current gold standard for trials of DSM-5 has been in agreement between two or more medical specialists. Research studies aiming to objectively predict depression have typically used brain scanning. Less expensive methods from cognitive neuroscience may allow quicker and more reliable diagnoses, and contribute to reducing the costs of managing the condition. In the current study we aimed to develop a novel inexpensive system for detecting elevated symptoms of depression based on tracking face and eye movements during the performance of cognitive tasks.
In total, 75 participants performed two novel cognitive tasks with verbal affective distraction elements while their face and eye movements were recorded using inexpensive cameras. Data from 48 participants (mean age 25.5 years, standard deviation of 6.1 years, 25 with elevated symptoms of depression) passed quality control and were included in a case-control classification analysis with machine learning.
Classification accuracy using cross-validation (within-study replication) reached 79% (sensitivity 76%, specificity 82%), when face and eye movement measures were combined. Symptomatic participants were characterised by less intense mouth and eyelid movements during different stages of the two tasks, and by differences in frequencies and durations of fixations on affectively salient distraction words.
Elevated symptoms of depression can be detected with face and eye movement tracking during the cognitive performance, with a close to clinically-relevant accuracy (~80%). Future studies should validate these results in larger samples and in clinical populations.
This study examines the use of discourse-level information to create expectations about reference in real-time processing, testing whether patterns previously observed among native speakers of English generalize to nonnative speakers. Findings from a visual-world eye-tracking experiment show that native (L1; N = 53) but not nonnative (L2; N = 52) listeners’ proactive coreference expectations are modulated by grammatical aspect in transfer-of-possession events. Results from an offline judgment task show these L2 participants did not differ from L1 speakers in their interpretation of aspect marking on transfer-of-possession predicates in English, indicating it is not lack of linguistic knowledge but utilization of this knowledge in real-time processing that distinguishes the groups. English proficiency, although varying substantially within the L2 group, did not modulate L2 listeners’ use of grammatical aspect for reference processing. These findings contribute to the broader endeavor of delineating the role of prediction in human language processing in general, and in the processing of discourse-level information among L2 users in particular.
The objective of the study was to evaluate the potential of Fourier transform infrared spectroscopy (FTIR) analysis of milk samples to predict body energy status and related traits (energy balance (EB), dry matter intake (DMI) and efficient energy intake (EEI)) in lactating dairy cows. The data included 2371 milk samples from 63 Norwegian Red dairy cows collected during the first 105 days in milk (DIM). To predict the body energy status traits, calibration models were developed using Partial Least Squares Regression (PLSR). Calibration models were established using split-sample (leave-one cow-out) cross-validation approach and validated using an external test set. The PLSR method was implemented using just the FTIR spectra or using the FTIR together with milk yield (MY) or concentrate intake (CONCTR) as predictors of traits. Analyses were conducted for the entire first 105 DIM and separately for the two lactation periods: 5 ≤ DIM ≤ 55 and 55 < DIM ≤ 105. To test the models, an external validation using an independent test set was performed. Predictions depending on the parity (1st, 2nd and 3rd-to 6th parities) in early lactation were also investigated. Accuracy of prediction (r) for both cross-validation and external test set was defined as the correlation between the predicted and observed values for body energy status traits. Analyzing FTIR in combination with MY by PLSR, resulted in relatively high r-values to estimate EB (r = 0.63), DMI (r = 0.83), EEI (r = 0.84) using an external validation. Only moderate correlations between FTIR spectra and traits like EB, EEI and dry matter intake (DMI) have so far been published. Our hypothesis was that improvements in the FTIR predictions of EB, EEI and DMI can be obtained by (1) stratification into different stages of lactations and different parities, or (2) by adding additional information on milking and feeding traits. Stratification of the lactation stages improved predictions compared with the analyses including all data 5 ≤ DIM ≤105. The accuracy was improved if additional data (MY or CONCTR) were included in the prediction model. Furthermore, stratification into parity groups, improved the predictions of body energy status. Our results show that FTIR spectral data combined with MY or CONCTR can be used to obtain improved estimation of body energy status compared to only using the FTIR spectra in Norwegian Red dairy cattle. The best prediction results were achieved using FTIR spectra together with MY for early lactation. The results obtained in the study suggest that the modeling approach used in this paper can be considered as a viable method for predicting an individual cow's energy status.
The objectives of this study is to predict the possible trajectory of coronavirus disease 2019 (COVID-19) spread in the United States. Prediction and severity ratings of COVID-19 are essential for pandemic control and economic reopening in the United States.
In this study, we apply the logistic and Gompertz model to evaluate possible turning points of the COVID-19 pandemic in different regions. By combining uncertainty and severity factors, this study constructed an indicator to assess the severity of the coronavirus outbreak in various states.
Based on the index of severity ratings, different regions of the United States are classified into 4 categories. The result shows that it is possible to identify the first turning point in Montana and Hawaii. It is unclear when the rest of the states will reach the first peak. However, it can be inferred that 75% of regions will not reach the first peak of coronavirus before August 2, 2020.
It is still essential for the majority of states to take proactive steps to fight against COVID-19 before August 2, 2020.
This article considers the role that assessment of suicidal ideation may have in short-term prediction of suicide. Suicide risk assessment is a multifactorial process and it is assumed that assessment of suicidal ideation is one component. Denial that suicidal ideation has any useful role in risk assessment fails to allow for the marked ongoing short-term variability in severity of intent, which is a common feature of the suicidal state of mind. It is concluded that the assessment of suicidal ideation, provided it is carried out correctly and applied appropriately, should continue to be regarded as a central component of the overall prediction process. A ‘two-take’ approach to short-term risk assessment and mitigation is proposed that takes variability in severity of intent into account and includes anticipatory treatment planning for any problems that may occur in the near future.
Given the fast spread of the novel coronavirus (COVID-19) worldwide and its classification by the World Health Organization (WHO) as being one of the worst pandemics in history, several scientific studies are carried out using various statistical and mathematical models to predict and study the likely evolution of this pandemic in the world. In the present research paper, we present a brief study aiming to predict the probability of reaching a new record number of COVID-19 cases in Lebanon, based on the record theory, giving more insights about the rate of its quick spread in Lebanon. The main advantage of the records theory resides in avoiding several statistical constraints concerning the choice of the underlying distribution and the quality of the residuals. In addition, this theory could be used, in cases where the number of available observations is somehow small. Moreover, this theory offers an alternative solution in case where machine learning techniques and long-term memory models are inapplicable because they need a considerable amount of data to be performant. The originality of this paper lies in presenting a new statistical approach allowing the early detection of unexpected phenomena such as the new pandemic COVID-19. For this purpose, we used epidemiological data from Johns Hopkins University to predict the trend of COVID-2019 in Lebanon. Our method is useful in calculating the probability of reaching a new record as well as studying the propagation of the disease. It also computes the probabilities of the waiting time to observe the future COVID-19 record. Our results obviously confirm the quick spread of the disease in Lebanon over a short time.
Why do humans make music? Theories of the evolution of musicality have focused mainly on the value of music for specific adaptive contexts such as mate selection, parental care, coalition signaling, and group cohesion. Synthesizing and extending previous proposals, we argue that social bonding is an overarching function that unifies all of these theories, and that musicality enabled social bonding at larger scales than grooming and other bonding mechanisms available in ancestral primate societies. We combine cross-disciplinary evidence from archaeology, anthropology, biology, musicology, psychology, and neuroscience into a unified framework that accounts for the biological and cultural evolution of music. We argue that the evolution of musicality involves gene-culture coevolution, through which proto-musical behaviors that initially arose and spread as cultural inventions had feedback effects on biological evolution due to their impact on social bonding. We emphasize the deep links between production, perception, prediction, and social reward arising from repetition, synchronization, and harmonization of rhythms and pitches, and summarize empirical evidence for these links at the levels of brain networks, physiological mechanisms, and behaviors across cultures and across species. Finally, we address potential criticisms and make testable predictions for future research, including neurobiological bases of musicality and relationships between human music, language, animal song, and other domains. The music and social bonding (MSB) hypothesis provides the most comprehensive theory to date of the biological and cultural evolution of music.
A Belgian predictive medical resource tool, Plan Risk Manifestations (PRIMA), for the prediction of the number of patient encounters at mass gatherings (MGs) has recently been developed, in addition to the existing models of Arbon and Hartman. This study presents the results of the validation process for the PRIMA model for music MGs.
A retrospective study was conducted using data gathered from music MGs in the province of Antwerp (Belgium) during the period of 2012-2016. Data from 87 music MGs were used for the study. The forecast of medical resources for these events was determined by entering the characteristics of individual events into the Arbon, Hartman, and PRIMA models. In order to determine if the PRIMA model is under- or over-predictive, the data gathered were retrospectively compared to the predicted number of resources needed using the aforementioned models. Statistical analysis included means, medians, and interquartile ranges (IQRs). Nonparametric related samples test (Wilcoxon Samples Signed Rank Test) for comparison of the median in deviations in predictions of patient presentation rates (PPRs) was performed using SPSS version 23 (IBM Corp.; Armonk, New York USA). Confidence interval levels were set at 95% and results were deemed statistically significant at P <.05. This triple comparison was used to determine the overall performance of all three models.
All three models had an acceptable rate of over-prediction of number of patient encounters ([Arbon 25.29%; 95% CI, 30.91-43.74]; [Hartman 29.89%; 95% CI, 57.10-68.90]; and [PRIMA 19.54%; 95% CI, 57.80-76.20]). But all models also had a high rate of under-prediction of number of patient encounters ([Arbon 74.71%; 95% CI, 453.31-752.52]; [Hartman 70.11%; 95% CI, 546.90-873.77]; and [PRIMA 78.16%; 95% CI, 288.91-464.89]). Only the PRIMA model succeeded in the correct prediction of the number of patient encounters on two occasions (2.3%).
Results of this study are in-line with existing literature. When comparing the predicted patient encounters, all three models had high rates of under-prediction and moderate rates of over-prediction. When comparing mean deviations, the PRIMA model had the lowest mean deviation of all predicted PPRs. Belgian events of the types included in the presented data may use the PRIMA model with confidence to predict PPRs and estimate the in-event health services (IEHS) requirements.
We present a statistical framework for generating predicted dynamic networks based on the observed evolution of social relationships in a population. The framework includes a novel and flexible procedure to sample dynamic networks given a probability distribution on evolving network properties; it permits the use of a broad class of approaches to model trends, seasonal variability, uncertainty, and changes in population composition. Current methods do not account for the variability in the observed historical networks when predicting the network structure; the proposed method provides a principled approach to incorporate uncertainty in prediction. This advance aids in the designing of network-based interventions, as development of such interventions often requires prediction of the network structure in the presence and absence of the intervention. Two simulation studies are conducted to demonstrate the usefulness of generating predicted networks when designing network-based interventions. The framework is also illustrated by investigating results of potential interventions on bill passage rates using a dynamic network that represents the sponsor/co-sponsor relationships among senators derived from bills introduced in the U.S. Senate from 2003 to 2016.
There is a need for accurate, inexpensive and field-friendly methods to assess body composition in children. Bioelectrical impedance analysis (BIA) is a promising approach; however, there have been limited validation and use among young children in resource-poor settings. We aim to develop and validate population-specific prediction equations for estimating total fat mass (FM), fat free-mass (FFM) and percentage body fat (PBF) in Vietnamese children (4–7 years) using reactance and resistance from BIA, anthropometric variables and demographic information. We conducted a cross-sectional survey of 120 children. Body composition was measured using dual-energy X-ray absorptiometry (DXA), BIA and anthropometry. To develop prediction equations, we split all data into development (70 %) and validation datasets (30 %). The model performance was evaluated using predicted residual error sum of squares, root mean squared error (RMSE), mean absolute error (MAE) and R2. We identified a top performing model with the least number of parameters (age, sex, weight and resistance index or resistance and height), low RMSE (FM 0·70, FFM 0·74, PBF 3·10), low MAE (FM 0·55, FFM 0·62, PBF 2·49), high R2 (FM 0·95, FFM 0·92, PBF 0·82) and the least difference between predicted values and actual values from DXA (FM 0·03 kg or 0·01 sd, FFM 0·06 kg or 0·02 sd, PBF 0·27 % or 0·04 sd). In conclusion, we developed the first valid and highly predictive equations to estimate FM, FFM and PBF in Vietnamese children using BIA. These findings have important implications for future research on the double burden of disease and risks associated with overweight and obesity in young children.
Phosphorus (P) is an essential nutrient in livestock feed but can pollute waterways. In order for pig production to become less of a threat to the environment, excreta must contain as little P as possible or be efficiently used by plants. This must be achieved without decreasing the livestock performance. Phosphorus and calcium (Ca) deposition in the bones of growing pigs must be optimised without affecting the muscle gain. This requires precision feeding based on cutting-edge techniques of diet formulation throughout the animal growth phase. Modelling and data mining have become important tools in this quest. In this study, a mechanistic model taking into account the distribution of P between bone and soft tissues was compared to the established factorial models (INRA (Jondreville and Dourmad, 2005) and NRC (National Research Council, 2012)) that predict P (apparent total tract digestible, ATTD-P; or standardised total tract digestible, STTD-P) and Ca (total and STTD) requirements as a function of BW and protein deposition. The requirements for different bone mineralisation scenarios, namely, 100% and 85% of the genetic potential, were compared with these two models. Sobol indices were used to estimate the relative impact of growth-related parameters on mineral requirements at 30, 60 and 120 kg of BW. The INRA showed the highest value of ATTD-P requirement between 29 and 103 kg of BW (6%) and lower for lighter and higher BW. Similarly, the model for 85% bone mineralisation showed lower STTD-P requirement than NRC between 29 and 93 kg of BW (7%) and higher for lighter and higher BW. Contrary to other models, the Ca requirement of the proposed model is not fixed in relation to P. It increases from 95 kg of BW while the others decrease. The INRA showed the highest Ca requirements. The model Ca requirements for 100% bone mineralisation are higher than NRC from 20 to 38 kg of BW similar until 70 kg of BW and then higher again. For 85% objective, the model showed lower Ca requirements from 25 to 82 kg of BW and higher for lighter and higher BW. The potential Ca deposition in bones is the most sensitive parameter (84% to 100% of the variance) of both ATTD-P and Ca at 30, 60 and 120 kg. The second most sensitive parameter is the protein deposition, explaining 1% to 15% of the ATTD-P variance. Studies such as this one will help to usher in a new era of sustainable and eco-friendly livestock production.
This study used event-related potentials (ERPs) to investigate how predicting upcoming words differ when contextual information used to generate the prediction is from the immediately preceding sentence context versus an earlier discourse context. Four-sentence discourses were presented to participants, with the critical words in the last sentences, either predictable or unpredictable based on sentence- or discourse-level contextual information. At the sentence level, the crucial contextual information for prediction was provided by the last sentence, where the critical word was embedded (e.g., Xiaoyu came to the living room. She made a cup of lemon tea. Then she sat down in a chair. She opened a box/an album to look at the pictures.), and at the discourse level by the first sentence (e.g., Xiaoyu took out a box/an album. She made a cup of lemon tea. Then she sat down in a chair. She leisurely looked at the pictures.). Results showed reduced N400 for predictable words compared to unpredictable counterparts at sentence and discourse levels and also a post-N400 positivity effect of predictability at sentence level. This suggests that both sentence- and discourse-level semantic information help readers predict upcoming words, but supportive sentence context more than discourse context.
To paraphrase the famous beginning of Douglas Adams’s The Hitchhiker’s Guide to the Galaxy: “Imagination is big. Really big. You just won’t believe how vastly hugely mindbogglingly big it is.” And it must be important, if even neuroscientists have noticed it. With momentum gathering to get a more comprehensive and interdisciplinary grip on the human imagination, at a time when still no one has much of a clue about this hypercomplex realm of the mind, we can expect to be treated to an initial proliferation of ideas, views, frameworks, and theories on the matter. But that is a good thing. There must be variation before there can be selection. Taking advantage of this temporarily heightened tolerance for speculation, this chapter proposes that the process of navigating an imaginary space can be described by different evolutionary algorithms that vary according to the prior knowledge we have of the topography of the imagined landscape or, put another way, the degree of sightedness we have of the fitness function of the imagined domain. Seen from this perspective, the category “novel combinatorial” of Abraham’s (2016) framework is best dissolved, with creative thinking being distributed and embedded into all other forms of the imagination.
Imagination – either explicitly or implicitly – plays an important role in contemporary conceptions of creativity. In contrast, imagination has not been given the same weight in most mainstream modern models of aesthetic experience. I argue that imagination is an important component of aesthetic experience in at least two ways. First, imagination likely guides our search for meaning when interacting with artworks. It can do so by driving our search for the underlying concepts and causes that originated the artwork, as well as facilitating internally generated thoughts. Second, imagination can facilitate transitions from states of uncertainty to states of increased predictability in the course of interacting with artworks. As such, models of aesthetic experience would benefit by explicitly incorporating imagination into their frameworks.
In this paper, we propose a multivariate Hawkes framework for modelling and predicting cyber attacks frequency. The inference is based on a public data set containing features of data breaches targeting the US industry. As a main output of this paper, we demonstrate the ability of Hawkes models to capture self-excitation and interactions of data breaches depending on their type and targets. In this setting, we detail prediction results providing the full joint distribution of future cyber attacks times of occurrence. In addition, we show that a non-instantaneous excitation in the multivariate Hawkes model, which is not the classical framework of the exponential kernel, better fits with our data. In an insurance framework, this study allows to determine quantiles for number of attacks, useful for an internal model, as well as the frequency component for a data breach guarantee.