A semi-supervised Bayesian mixture modelling approach for joint batch correction and classification

Systematic differences between batches of samples present significant challenges when analysing biological data. Such batch effects are well-studied and are liable to occur in any setting where multiple batches are assayed. Many existing methods for accounting for these have focused on high-dimensional data such as RNA-seq and have assumptions that reflect this. Here we focus on batch-correction in low-dimensional classification problems. We propose a semi-supervised Bayesian generative classifier based on mixture models that jointly predicts class labels and models batch effects. Our model allows observations to be probabilistically assigned to classes in a way that incorporates uncertainty arising from batch effects. We explore two choices for the within-class densities: the multivariate normal and the multivariate t. A simulation study demonstrates that our method performs well compared to popular off-the-shelf machine learning methods and is also quick; performing 15,000 iterations on a dataset of 500 samples with 2 measurements each in 7.3 seconds for the MVN mixture model and 11.9 seconds for the MVT mixture model. We apply our model to two datasets generated using the enzyme-linked immunosorbent assay (ELISA), a spectrophotometric assay often used to screen for antibodies. The examples we consider were collected in 2020 and measure seropositivity for SARS-CoV-2. We use our model to estimate seroprevalence in the populations studied. We implement the models in C++ using a Metropolis-within-Gibbs algorithm; this is available in the R package at https://github.com/stcolema/BatchMixtureModel. Scripts to recreate our analysis are at https://github.com/stcolema/BatchClassifierPaper.

A self-adaptive level-based learning artificial bee colony algorithm for feature selection on high-dimensional classification

Feature selection is an important data preprocessing method in data mining and machine learning, yet it faces the challenge of “curse of dimensionality” when dealing with high-dimensional data. In this paper, a self-adaptive level-based learning artificial bee colony (SLLABC) algorithm is proposed for high-dimensional feature selection problem. The SLLABC algorithm includes three new mechanisms: (1) A novel level-based learning mechanism is introduced to accelerate the convergence of the basic artificial bee colony algorithm, which divides the population into several levels and the individuals on each level learn from the individuals on higher levels, especially, the individuals on the highest level learn from each other. (2) A self-adaptive method is proposed to keep the balance between exploration and exploitation abilities, which takes the diversity of population into account to determine the number of levels. The lower the diversity is, the fewer the levels are divided. (3) A new update mechanism is proposed to reduce the number of selected features. In this mechanism, if the error rate of an offspring is higher than or is equal to that of its parent but selects more features, then the offspring is discarded and the parent is retained, otherwise, the offspring replaces its parent. Further, we discuss and analyze the contribution of these novelties to the diversity of population and the performance of classification. Finally, the results, compared with 8 state-of-the-art algorithms on 12 high-dimensional datasets, confirm the competitive performance of the proposed SLLABC on both classification accuracy and the size of the feature subset.

Energy-Oriented Production Planning in Industry: A Systematic Literature Review and Classification Scheme

Scarcity of resources, structural change during the further development of renewable energy sources, and their corresponding costs, such as increasing resource costs or penalties due to dirty production, lead industrial firms to adapt ecological actions. In this regard, research on energy utilization in production planning has received increased attention in the last years, resulting in a large number of research articles so far. With the paper at hand, we review the literature on energy-oriented production planning. The aim of this study is to derive similar core issues and related properties along energy-oriented models within hierarchical production planning. For this, we carry out a systematic literature review and analyze and synthesize 375 research articles. We classify the underlying literature with a novel two-dimensional classification scheme and identify three key topics and five frequently found characteristics, which are presented in detail throughout this article. Based on these results, we state several potentials for further research.

Development of Academic Technological Entrepreneurship in Russia

Technology entrepreneurs play a critical role in modern models of economic growth. At the same time, the features and development trajectories of technological entrepreneurship and gazelle firms in Russia differ from countries with mature market economies. The purpose of the article is to assess the state of academic technology entrepreneurship in Russian universities and to develop recommendations for its development. The methods of two-dimensional classification of universities according to two indicators of technological entrepreneurship development, correlation analysis, assessment of intergroup differences according to the Mann-Whitney U-criterion, qualitative analysis of entrepreneurial ecosystems in universities, methods of constructing algorithms were used. As a result, it was found that the number of start-ups and the likelihood of receiving commercial funding are practically not correlated. There is a group of universities that are not among the largest metropolitan universities, but have a high proportion of entrepreneurial projects that have successfully passed the seed stage and received commercial funding (business angel, venture fund). These top performing universities are distinguished not only by their innovative infrastructure, but also by a wide variety of community centers. Taking into account the results obtained an algorithm for the development of academic technological entrepreneurship on the basis of the university as an innovation hub has been developed. Its important elements are: pre-active marketing of scientific research groundwork, development of the social capital of the ecosystem, and collaboration practices. The results may be of interest to technology entrepreneurs themselves, as well as to universities interested in the development of academic entrepreneurship.

Testing extra-linearity across a psychosis continuum

Background It is unclear whether psychotic experiences (PEs) gradually merge into states of clinical psychosis along a continuum which correspond to a dimensional classification or whether latent classes appear above a certain severity threshold which correspond better to diagnostic categories of psychosis. Methods Annual cross-sectional surveys, 2014–19, among Chinese undergraduates (N = 47,004) measured PEs, depression and etiological risk factors using standardized self-report instruments. We created a psychosis continuum with five levels and tested linear and extra-linear contrasts in associated etiological risk factors, before and after adjustment for depression. We carried out latent class analysis. Results Categorical expression of psychosis, including hallucinations and delusions, nuclear symptoms, and nuclear symptoms and depression were found at severe level 5. Etiological risk factors which impacted linearly across the continuum were more common for depression. Child maltreatment impacted extra-linearly on both psychosis and depression. Family history of psychosis impacted linearly on psychosis; male sex and urban birth impacted extra-linearly and were specific for psychosis. Four latent classes were found, but only at level 5. These corresponded to nuclear schizophrenia symptoms, nuclear schizophrenia and depressive symptoms, severe depression, and an unclassified category with moderate prevalence of PEs. Conclusion Quantitative and qualitative changes in the underlying structure of psychosis were observed at the most severe level along a psychosis continuum, where four latent classes emerged. These corresponded to existing categorical classifications but require confirmation with clinical interview. PEs are non-specific and our findings suggest some are on a continuum with depression, whilst others are on a continuum with non-affective psychosis. Differing patterns of impact from etiological risk factors across the spectrum of psychopathology determine outcome at the most severe level of these continua.

The Effect of Emotion on Time Perception for Complex Visual Stimuli

<p>Activation and attention have opposite effects on time perception. Emotion can both increase physiological activation (which leads to overestimation of time) and attract attention (which leads to underestimation of time). Although the effect of emotion on time perception has received a growing amount of attention, the use of different time estimation tasks and stimuli makes it difficult to compare findings across studies. The effect of emotion on the temporal perception of complex stimuli (e.g. scenes) is particularly under-researched. This thesis presents a systematic assessment of the effect of two key emotional dimensions, arousal and valence, on time perception for visual stimuli. Studies were designed to control for factors that may modulate emotion effects, such as image repetition and carry over from one emotion to another. The stimuli were complex images standardized for arousal (high or low) and valence (positive or negative) as well as neutral images. The first study compared three time estimation tasks to determine which were sensitive to emotion effects. The selected task, temporal bisection, was used to test time perception in three duration ranges: short (400 to 1600ms), middle (1000 to 4000ms), and long (2000 to 6000ms). Results of bisection point analyses revealed that the duration of attention-capturing stimuli (e.g. high arousal or negative images) was underestimated compared to that of other stimuli (e.g. low arousal or neutral images). These findings are at odds with activational effects of emotion (overestimation of emotional stimuli), which are typically found in studies of time perception for facial expression. Better temporal sensitivity in the long range than in short and middle ranges suggests that participants used different timing strategies to perform the bisection task at longer stimulus durations. To test the effect of emotion on time perception using a discrete rather than dimensional classification of emotion, experiments were replicated using emotional facial expressions as stimuli. Time estimates in the short and middle ranges did not show attentional effects, but pointed to activational effects of emotion. Facial expression had no impact on time perception in the long duration range. Taken together, these experiments show that the effect of emotion on time perception varies according to both duration and stimulus type. Emotional facial expressions have short lived activational effects whereby the duration of arousing stimuli is overestimated, whereas complex emotional scenes have protracted attentional effects through which the duration of attention-capturing stimuli is underestimated.</p>

The Effect of Emotion on Time Perception for Complex Visual Stimuli

<p>Activation and attention have opposite effects on time perception. Emotion can both increase physiological activation (which leads to overestimation of time) and attract attention (which leads to underestimation of time). Although the effect of emotion on time perception has received a growing amount of attention, the use of different time estimation tasks and stimuli makes it difficult to compare findings across studies. The effect of emotion on the temporal perception of complex stimuli (e.g. scenes) is particularly under-researched. This thesis presents a systematic assessment of the effect of two key emotional dimensions, arousal and valence, on time perception for visual stimuli. Studies were designed to control for factors that may modulate emotion effects, such as image repetition and carry over from one emotion to another. The stimuli were complex images standardized for arousal (high or low) and valence (positive or negative) as well as neutral images. The first study compared three time estimation tasks to determine which were sensitive to emotion effects. The selected task, temporal bisection, was used to test time perception in three duration ranges: short (400 to 1600ms), middle (1000 to 4000ms), and long (2000 to 6000ms). Results of bisection point analyses revealed that the duration of attention-capturing stimuli (e.g. high arousal or negative images) was underestimated compared to that of other stimuli (e.g. low arousal or neutral images). These findings are at odds with activational effects of emotion (overestimation of emotional stimuli), which are typically found in studies of time perception for facial expression. Better temporal sensitivity in the long range than in short and middle ranges suggests that participants used different timing strategies to perform the bisection task at longer stimulus durations. To test the effect of emotion on time perception using a discrete rather than dimensional classification of emotion, experiments were replicated using emotional facial expressions as stimuli. Time estimates in the short and middle ranges did not show attentional effects, but pointed to activational effects of emotion. Facial expression had no impact on time perception in the long duration range. Taken together, these experiments show that the effect of emotion on time perception varies according to both duration and stimulus type. Emotional facial expressions have short lived activational effects whereby the duration of arousing stimuli is overestimated, whereas complex emotional scenes have protracted attentional effects through which the duration of attention-capturing stimuli is underestimated.</p>

Implementation of innovative medical technologies in German inpatient care: patterns of utilization and evidence development

Background Innovative medical technologies are commonly associated with positive expectations. At the time of their introduction into care, there is often little evidence available regarding their benefits and harms. Accordingly, some innovative medical technologies with a lack of evidence are used widely until or even though findings of adverse events emerge, while others with study results supporting their safety and effectiveness remain underused. This study aims at examining the diffusion patterns of innovative medical technologies in German inpatient care between 2005 and 2017 while simultaneously considering evidence development. Methods Based on a qualitatively derived typology and a quantitative clustering of the adoption curves, a representative sample of 21 technologies was selected for further evaluation. Published scientific evidence on efficacy/effectiveness and safety of the technologies was identified and extracted in a systematic approach. Derived from a two-dimensional classification according to the degree of utilization and availability of supportive evidence, the diffusion patterns were then assigned to the categories “Success” (widespread/positive), “Hazard” (widespread/negative), “Overadoption” (widespread/limited or none), “Underadoption” (cautious/positive), “Vigilance” (cautious/negative), and “Prudence” (cautious/limited or none). Results Overall, we found limited evidence on the examined technologies regarding both the quantity and quality of published randomized controlled trials. Thus, the categories “Prudence” and “Overadoption” together account for nearly three-quarters of the years evaluated, followed by “Success” with 17%. Even when evidence is available, the transfer of knowledge into practice appears to be inhibited. Conclusions The successful implementation of safe and effective innovative medical technologies into practice requires substantial further efforts by policymakers to strengthen systematic knowledge generation and translation. Creating an environment that encourages the conduct of rigorous studies, promotes knowledge translation, and rewards innovative medical technologies according to their added value is a prerequisite for the diffusion of valuable health care.