Semi-supervised Machine Learning Approach Research Articles

Identifying type II quasars at intermediate redshift with few-shot learning photometric classification

A sub-population of AGNs where the central engine is obscured are known as type II quasars (QSO2s). These luminous AGNs have a thick and dusty torus that obscures the accretion disc from our line of sight. Thus, their special orientation allows for detailed studies of the AGN-host co-evolution. Increasing the sample size of QSO2 sources in critical redshift ranges is crucial for understanding the interplay of AGN feedback, the AGN-host relationship, and the evolution of active galaxies. We aim to identify QSO2 candidates in the `redshift desert' using optical and infrared photometry. At this intermediate redshift range (i.e. $1 z 2$), most of the prominent optical emission lines in QSO2 sources (e.g. CIV$ OIII 4959, 5008$) fall either outside the wavelength range of the SDSS optical spectra or in particularly noisy wavelength ranges, making QSO2 identification challenging. Therefore, we adopted a semi-supervised machine learning approach to select candidates in the SDSS galaxy sample. Recent applications of machine learning in astronomy focus on problems involving large data sets, with small data sets often being overlooked. We developed a `few-shot' learning approach for the identification and classification of rare-object classes using limited training data (200 sources). The new AMELIA pipeline uses a transfer-learning based approach with decision trees, distance-based, and deep learning methods to build a classifier capable of identifying rare objects on the basis of an observational training data set. We validated the performance of AMELIA by addressing the problem of identifying QSO2s at $1 z 2$ using SDSS and WISE photometry, obtaining an F1-score above 0.8 in a supervised approach. We then used AMELIA to select new QSO2 candidates in the `redshift desert' and examined the nature of the candidates using SDSS spectra, when available. In particular, we identified a sub-population of NeV emitters at $z which are highly likely to contain obscured AGNs. We used X-ray and radio cross-matching to validate our classification and investigated the performance of photometric criteria from the literature showing that our candidates have an inherent dusty nature. Finally, we derived physical properties for our QSO2 sample using photoionisation models and verified the AGN classification using an SED fitting. Our results demonstrate the potential of few-shot learning applied to small data sets of rare objects, in particular QSO2s, and confirms that optical-IR information can be further explored to search for obscured AGNs. We present a new sample of candidates to be further studied and validated using multi-wavelength observations.

Semi-supervised machine learning approach for reaction stoichiometry and kinetic model identification using spectral data from flow reactors

The semi-supervised machine learning approach is an integrated calibration-free modelling framework for identifying reaction systems from spectral data using minimal prior information and it is validated with experimental data obtained in a micro-reactor.

Semi-Supervised Machine Learning Approaches for DDOS Attack Detection

Network infrastructures are the target of several attacks. These include intrusions into the confidentiality, integrity, and availability of the network. The network's availability is impacted by a persistent attack known as a distributed denial-of-service (DDoS) attack. Such an assault is carried out using a command and control (C & C) technique. To detect these assaults, numerous researchers have put forth various machine learning-based solutions. In this paper, we are going to detect different DDoS attacks by various methods and evaluate their performance. This experiment made use of the KD99 dataset. The normal and assault samples were classified using the random forest technique. The classification of 99.76% of the samples was accurate. By strategically selecting clusters and incorporating the insights gained from the small labelled dataset, a portion of the unlabelled clusters can be assigned labels, effectively converting raw data into useful training examples. This enriched dataset is then used to train an improved classifier that can better generalize and adapt to the dynamic nature of DDoS attacks.

A Machine Learning Approach to Predicting Physical Activity Levels in Adolescents

The ongoing evolution of technology has had both positive and negative effects on modern society. On the positive side, it has significantly improved the ease with which various activities can be performed. However, it has also had a negative impact by reducing physical activity. This reduction in physical activity, in turn, increases the risk of chronic diseases that contribute to global mortality rates. This research aims to assess the effectiveness of machine learning in predicting the physical activity levels of adolescents. The study utilizes data from accelerometers, specifically the ActiGraph GT3X. The research methodology employs a semi-supervised machine learning approach, using the support vector machine and decision tree algorithms to make these predictions. The sample comprises 61 adolescents (males = 17, female = 44), including high school students and university students aged 18-21, from the West Java region. The results from the machine learning model using the decision tree algorithm indicated a model accuracy of 97.50% in predicting physical activity levels. In contrast, the accuracy obtained from the performance analysis using the confusion matrix for the support vector machine model was 92.5%. Based on these accuracy levels, the decision tree algorithm outperforms the support vector machine algorithm's accuracy. Further analyses involving different models are necessary to determine which algorithm offers the highest level of accuracy.

Comparing supervised and semi-supervised machine learning approaches in NTCP modeling to predict complications in head and neck cancer patients

Background and purposeHead and neck cancer (HNC) patients treated with radiotherapy often suffer from radiation-induced toxicities. Normal Tissue Complication Probability (NTCP) modeling can be used to determine the probability to develop these toxicities based on patient, tumor, treatment and dose characteristics. Since the currently used NTCP models are developed using supervised methods that discard unlabeled patient data, we assessed whether the addition of unlabeled patient data by using semi-supervised modeling would gain predictive performance. Materials and methodsThe semi-supervised method of self-training was compared to supervised regression methods with and without prior multiple imputation by chained equation (MICE). The models were developed for the most common toxicity outcomes in HNC patients, xerostomia (dry mouth) and dysphagia (difficulty swallowing), measured at six months after treatment, in a development cohort of 750 HNC patients. The models were externally validated in a validation cohort of 395 HNC patients. Model performance was assessed by discrimination and calibration. ResultsMICE and self-training did not improve performance in terms of discrimination or calibration at external validation compared to current regression models. In addition, the relative performance of the different models did not change upon a decrease in the amount of (labeled) data available for model development. Models using ridge regression outperformed the logistic models for the dysphagia outcome. ConclusionSince there was no apparent gain in the addition of unlabeled patient data by using the semi-supervised method of self-training or MICE, the supervised regression models would still be preferred in current NTCP modeling for HNC patients.

On Evaluating IoT Data Trust via Machine Learning

Data trust in IoT is crucial for safeguarding privacy, security, reliable decision-making, user acceptance, and complying with regulations. Various approaches based on supervised or unsupervised machine learning (ML) have recently been proposed for evaluating IoT data trust. However, assessing their real-world efficacy is hard mainly due to the lack of related publicly available datasets that can be used for benchmarking. Since obtaining such datasets is challenging, we propose a data synthesis method, called random walk infilling (RWI), to augment IoT time-series datasets by synthesizing untrustworthy data from existing trustworthy data. Thus, RWI enables us to create labeled datasets that can be used to develop and validate ML models for IoT data trust evaluation. We also extract new features from IoT time-series sensor data that effectively capture its autocorrelation as well as its cross-correlation with the data of the neighboring (peer) sensors. These features can be used to learn ML models for recognizing the trustworthiness of IoT sensor data. Equipped with our synthesized ground-truth-labeled datasets and informative correlation-based features, we conduct extensive experiments to critically examine various approaches to evaluating IoT data trust via ML. The results reveal that commonly used ML-based approaches to IoT data trust evaluation, which rely on unsupervised cluster analysis to assign trust labels to unlabeled data, perform poorly. This poor performance is due to the underlying assumption that clustering provides reliable labels for data trust, which is found to be untenable. The results also indicate that ML models, when trained on datasets augmented via RWI and using the proposed features, generalize well to unseen data and surpass existing related approaches. Moreover, we observe that a semi-supervised ML approach that requires only about 10% of the data labeled offers competitive performance while being practically more appealing compared to the fully supervised approaches. The related Python code and data are available online.

A machine learning-based score for precise echocardiographic assessment of cardiac remodelling in hypertensive young adults.

Accurate staging of hypertension-related cardiac changes, before the development of significant left ventricular hypertrophy, could help guide early prevention advice. We evaluated whether a novel semi-supervised machine learning approach could generate a clinically meaningful summary score of cardiac remodelling in hypertension. A contrastive trajectories inference approach was applied to data collected from three UK studies of young adults. Low-dimensional variance was identified in 66 echocardiography variables from participants with hypertension (systolic ≥160 mmHg) relative to a normotensive group (systolic < 120 mmHg) using a contrasted principal component analysis. A minimum spanning tree was constructed to derive a normalized score for each individual reflecting extent of cardiac remodelling between zero (health) and one (disease). Model stability and clinical interpretability were evaluated as well as modifiability in response to a 16-week exercise intervention. A total of 411 young adults (29 ± 6 years) were included in the analysis, and, after contrastive dimensionality reduction, 21 variables characterized >80% of data variance. Repeated scores for an individual in cross-validation were stable (root mean squared deviation = 0.1 ± 0.002) with good differentiation of normotensive and hypertensive individuals (area under the receiver operating characteristics 0.98). The derived score followed expected hypertension-related patterns in individual cardiac parameters at baseline and reduced after exercise, proportional to intervention compliance (P = 0.04) and improvement in ventilatory threshold (P = 0.01). A quantitative score that summarizes hypertension-related cardiac remodelling in young adults can be generated from a computational model. This score might allow more personalized early prevention advice, but further evaluation of clinical applicability is required.

Solo tourism: exploration and conceptualization – a semi-supervised machine learning approach

ABSTRACT This study aims at understanding and conceptualizing solo tourism. By using a semi-supervised machine learning approach and scrutinizing 27,208 solo tourist tweets. Based on implicit self-theory we capture solo consumers’ self-development and self-discovery. We provide an all-inclusive slow tourism conceptualization and show that the solo tourism framework is a three-stage journey of self-discovery. This study not only provides tourism scholars and providers with an evidence-based solo tourism conceptualization, but also with a marketing, psychological, and operation tool to manage this segment.

Automation of the Timed Up and Go Test Using a Doppler Radar System for Gait and Balance Analysis in Elderly People.

The timed up and go (TUG) test is a simple, valid, and reliable clinical tool that is widely used to assess mobility in elderly people. Several research studies have been conducted to automate the TUG test using wearable sensors or motion-tracking systems. Despite their promising results, the adopted technological systems present inconveniences in terms of acceptability and privacy protection. In this work, we propose to overcome these problems by using a Doppler radar system set into the backrest of a chair in order to automate the TUG test and extract additional information from its phases (i.e., transfer, walk, and turn). We intend to segment its phases and extract spatiotemporal gait parameters automatically. Our methodology is mainly based on a multiresolution analysis of radar signals. We proposed a segmentation technique based on the extraction of limbs oscillations signals through a semisupervised machine learning approach, on the one hand, and the application of the DARC algorithm on the other hand. Once the speed signals of torso and limbs oscillations were detected, we suggested estimating 14 gait parameters. All our approaches were validated by comparing outcomes to those obtained from a reference Vicon system. High correlation coefficients were obtained by comparing the speed signals of the torso (ρ=0.8), the speed signals of limbs oscillations (ρ=0.91), the initial and final indices of TUG phases (ρ=0.95), and the extracted parameters (percentage error < 4.8) obtained after radar signal processing to those obtained from the Vicon system.

Abstract 650: A machine learning model accurately predicts response to immune checkpoint inhibitors in colorectal cancer using the gut microbiome

Abstract Background: Immune checkpoint inhibitors (ICI) have shown a long-lasting response in many cancers such as colorectal cancer. However, only patients with microsatellite instability (MSI-High) or mismatch repair deficiency (dMMR) benefit from ICI. One of the many factors that can influence ICI-response is the gut microbiome, however its effect remains elusive. Here we aim to investigate the role of the gut microbiome in ICI-response using machine learning (ML). Methods: Clinical and genomic data of ICI-treated CRC patients were retrieved from the KEYNOTE-177 trial. Microbiome abundances from The Cancer Genome Atlas (TCGA) were downloaded from the cBioportal database. A label propagation semi-supervised ML approach was conducted to label patients’ responses to ICI treatment in the TCGA cohort. Tumor mutational burden (TMB), MSI status, treatment, histological subtype, TNM staging and prior chemotherapy were fitted into the semi-supervised model to predict ICI-response in the combination cohort. The normalized abundance of 1406 microbial signatures were regularized using a LASSO model to avoid multicollinearity. Signatures with a non-zero LASSO coefficients were fitted into a supervised ML Random Forest Classification model (RFC) to evaluate their prediction for ICI-response. Samples were split with 80:20 training-testing ratio, and model’s performance was evaluated on the testing set using mean bootstrap estimate, 10-fold cross-validation (CV), and area under curve (AUC). Results: A total of 538 CRC patients from the TCGA and 29 ICI-treat patients from the KEYNOTE-177 trial were fitted into the semi-supervised model to label ICI-response in the TCGA cohort. A total of 95 patients were labeled as responder (R) while 488 patients were labeled as non-responder (NR). Responders had significantly higher TMB levels than NR (mean TMB: 22.5 vs. 12.2, p-value &amp;lt;0.0001). Of the 95 R, 42 were MSS, and 53 patients were MSI-High. While 417 of the NR were MSS. A total of 25 microbial signatures were screened after LASSO for their ICI prediction and were fitted into the RFC model along with TMB and MSI status (Hereafter called RFC27). The RFC27 model performed on the testing set with mean bootstrap estimate of 0.88 and 95%CI: [0.83-0.94], 10-fold CV of 0.84 and AUC of 0.97. TMB and MSI showed highest feature contribution followed by Actinopolyspora and Bacteroides Eggerthia. The relative abundance of both microbiotas was significantly different between R and NR (mean: 0.99 vs. 0.47, 2.0 vs. 1.6; p-value &amp;lt;0.001 respectively). Conclusions: Our results show that the gut microbiome can accurately predict ICI-response and their higher abundance contributed to a better response. Previous studies also reported a pro-inflammatory response of Bacteroides in addition to their association with ICI-response in melanoma patients, thus providing a potential therapeutic marker for ICI. Citation Format: Ayah N. Al-Bzour. A machine learning model accurately predicts response to immune checkpoint inhibitors in colorectal cancer using the gut microbiome [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2023; Part 1 (Regular and Invited Abstracts); 2023 Apr 14-19; Orlando, FL. Philadelphia (PA): AACR; Cancer Res 2023;83(7_Suppl):Abstract nr 650.

Automatic classification of construction safety reports using semi-supervised YAKE-Guided LDA approach

Literature on supervised Machine-Learning (ML) approaches for classifying text-based safety reports for the construction sector has been growing. Recent studies have emphasized the need to build ML approaches that balance high classification accuracy and performance on management criteria, such as resource intensiveness. However, despite being highly accurate, the extensively focused, supervised ML approaches may not perform well on management criteria as many factors contribute to their resource intensiveness. Alternatively, the potential for semi-supervised ML approaches to achieve balanced performance has rarely been explored in the construction safety literature. The current study contributes to the scarce knowledge on semi-supervised ML approaches by demonstrating the applicability of a state-of-the-art semi-supervised learning approach, i.e., Yet, Another Keyword Extractor (YAKE) integrated with Guided Latent Dirichlet Allocation (GLDA) for construction safety report classification. Construction-safety-specific knowledge is extracted as keywords through YAKE, relying on accessible literature with minimal manual intervention. Keywords from YAKE are then seeded in the GLDA model for the automatic classification of safety reports without requiring a large quantity of prelabeled datasets. The YAKE-GLDA classification performance (F1 score of 0.66) is superior to existing unsupervised methods for the benchmark data containing injury narratives from Occupational Health and Safety Administration (OSHA). The YAKE-GLDA approach is also applied to near-miss safety reports from a construction site. The study demonstrates a high degree of generality of the YAKE-GLDA approach through a moderately high F1 score of 0.86 for a few categories in the near-miss data. The current research demonstrates that, unlike the existing supervised approaches, the semi-supervised YAKE-GLDA approach can achieve a novel possibility of consistently achieving reasonably good classification performance across various construction-specific safety datasets yet being resource-efficient. Results from an objective comparative and sensitivity analysis contribute to much-required knowledge-contesting insights into the functioning and applicability of the YAKE-GLDA. The results from the current study will help construction organizations implement and optimize an efficient ML-based knowledge-mining strategy for domains beyond safety and across sites where the availability of a pre-labeled dataset is a significant limitation.

Leveraging Active Learning for Failure Mode Acquisition

Identifying failure modes is an important task to improve the design and reliability of a product and can also serve as a key input in sensor selection for predictive maintenance. Failure mode acquisition typically relies on experts or simulations which require significant computing resources. With the recent advances in Natural Language Processing (NLP), efforts have been made to automate this process. However, it is not only time consuming, but extremely challenging to obtain maintenance records that list failure modes. Unsupervised learning methods such as topic modeling, clustering, and community detection are promising approaches for automatic processing of maintenance records to identify failure modes. However, the nascent state of NLP tools combined with incompleteness and inaccuracies of typical maintenance records pose significant technical challenges. As a step towards addressing these challenges, this paper proposes a framework in which online active learning is used to identify failure modes from maintenance records. Active learning provides a semi-supervised machine learning approach, allowing for a human in the training stage of the model. The hypothesis of this paper is that the use of a human to annotate part of the data and train a machine learning model to annotate the rest is more efficient than training unsupervised learning models. Results demonstrate that the model is trained with annotating less than ten percent of the total available data. The framework is able to achieve ninety percent (90%) accuracy in the identification of failure modes in test cases with an F-1 score of 0.89. This paper also demonstrates the effectiveness of the proposed framework with both qualitative and quantitative measures.

Biphasic burst and sustained transdermal delivery in vivo using an AI-optimized 3D-printed MN patch

The objective of the present study was to fabricate microneedles for delivering lipophilic active ingredients (APIs) using digital light processing (DLP) printing technology and quality by design (QbD) supplemented by artificial intelligence (AI) algorithms. In the present study, dissolvable microneedle (MN) patches using ibuprofen (IBU) as a model drug were successfully fabricated with DLP printing technology at ∼ 750 μm height, ∼250 μm base diameter, and tip with radius of curvature (RoC) of ∼ 15 μm. MN patches were comprised of IBU, photoinitiator, Lithium phenyl (2,4,6-trimethylbenzoyl) phosphinate (LAP), polyethylene glycol dimethacrylate (PEGDAMA)550 and distilled water and were developed using the QbD optimization approach. Optimization of print fidelity and needle morphology were achieved using AI implementing a semi-supervised machine learning approach. Mechanical strength tests demonstrated that IBU MNs formed pores both on Parafilm M® and human cadaver skin. IBU-MNs consisting of 0.23 %w/v and 0.49 %w/v LAP with 10 %w/v water showed ∼ 2 mg/cm2 sustained drug permeation at 72 h in skin permeation experiments with flux of ∼ 40 μg/cm2/h. Pharmacokinetic studies in rats displayed biphasic rapid first-order absorption with sustained zero-order input of Ko = 150ug/hr, AUC0-48h = 62812.02 ± 11128.39 ng/ml*h, Tmax = 2.66 ± 1.12 h, and Cmax = 3717.43 ± 782.25 ng/ml (using 0.23 %w/v LAP IBU MN patch). An in vitro in vivo relation (IVIVR) was conducted identifying a polynomial relationship between patch release and fraction absorbed in vivo. This study demonstrates fabrication of dissolvable DLP-printed microneedle patches for lipophilic API delivery with biphasic rapid first-order and sustained zero-order release.

Machine learning-driven clinical decision support system for concept-based searching: a field trial in a Norwegian hospital

BackgroundNatural language processing (NLP) based clinical decision support systems (CDSSs) have demonstrated the ability to extract vital information from patient electronic health records (EHRs) to facilitate important decision support tasks. While obtaining accurate, medical domain interpretable results is crucial, it is demanding because real-world EHRs contain many inconsistencies and inaccuracies. Further, testing of such machine learning-based systems in clinical practice has received limited attention and are yet to be accepted by clinicians for regular use.MethodsWe present our results from the evaluation of an NLP-driven CDSS developed and implemented in a Norwegian Hospital. The system incorporates unsupervised and supervised machine learning combined with rule-based algorithms for clinical concept-based searching to identify and classify allergies of concern for anesthesia and intensive care. The system also implements a semi-supervised machine learning approach to automatically annotate medical concepts in the narrative.ResultsEvaluation of system adoption was performed by a mixed methods approach applying The Unified Theory of Acceptance and Use of Technology (UTAUT) as a theoretical lens. Most of the respondents demonstrated a high degree of system acceptance and expressed a positive attitude towards the system in general and intention to use the system in the future. Increased detection of patient allergies, and thus improved quality of practice and patient safety during surgery or ICU stays, was perceived as the most important advantage of the system.ConclusionsOur combined machine learning and rule-based approach benefits system performance, efficiency, and interpretability. The results demonstrate that the proposed CDSS increases detection of patient allergies, and that the system received high-level acceptance by the clinicians using it. Useful recommendations for further system improvements and implementation initiatives are reducing the quantity of alarms, expansion of the system to include more clinical concepts, closer EHR system integration, and more workstations available at point of care.

Estimating the Prevalence of Dementia in India Using a Semi-Supervised Machine Learning Approach

Introduction: Accurate estimation of dementia prevalence is essential for making effective public and social care policy to support individuals and families suffering from the disease. The purpose of this paper is to estimate the prevalence of dementia in India using a semi-supervised machine learning approach based on a large nationally representative sample. Methods: The sample of this study is adults 60 years or older in the wave 1 (2017–2019) of the Longitudinal Aging Study in India (LASI). A subsample in LASI received extensive cognitive assessment and clinical consensus ratings and therefore has diagnoses of dementia. A semi-supervised machine learning model was developed to predict the status of dementia for LASI participants without diagnoses. After obtaining the predictions, sampling weights and age standardization to the World Health Organization (WHO) standard population were applied to generate the estimate for prevalence of dementia in India. Results: The prevalence of dementia for those aged 60 years and older in India was 8.44% (95% CI: 7.89%–9.01%). The age-standardized prevalence was estimated to be 8.94% (95% CI: 8.36%–9.55%). The prevalence of dementia was greater for those who were older, were females, received no education, and lived in rural areas. Discussion: The prevalence of dementia in India may be higher than prior estimates derived from local studies. These prevalence estimates provide the information necessary for making long-term planning of public and social care policy. The semi-supervised machine learning approach adopted in this paper may also be useful for other large population aging studies that have a similar data structure.

Machine learning analysis to classify nanoparticles from noisy spICP-TOFMS data

A two-stage semi-supervised machine learning approach was developed as a robust method to classify cerium-rich engineered, incidental, and natural nanoparticles measured by spICP-TOFMS.

A Semi-Supervised Machine Learning Approach in Predicting High-Risk Pregnancies in the Philippines.

Early risk tagging is crucial in maternal health, especially because it threatens both the mother and the long-term development of the baby. By tagging high-risk pregnancies, mothers would be given extra care before, during, and after pregnancies, thus reducing the risk of complications. In the Philippines, where the fertility rate is high, especially among the youth, awareness of risks can significantly contribute to the overall outcome of the pregnancy and, to an extent, the Maternal mortality rate. Although supervised machine learning models have ubiquity as predictors, there is a gap when data are weak or scarce. Using limited collected data from the municipality of Daraga in Albay, the study first compared multiple supervised machine learning algorithms to analyze and accurately predict high-risk pregnancies. Through hyperparameter tuning, supervised learning algorithms such as Decision Tree, Random Forest, Support Vector Machine, K-Nearest Neighbors, Naïve Bayes, and Multilayer Perceptron were evaluated by using 10-fold cross validation to obtain the best parameters with the best scores. The results show that Decision Tree bested other algorithms and attained a test score of 93.70%. To address the gap, a semi-supervised approach using a Self-Training model was applied to the modified Decision Tree, which was then used as the base estimator with a 30% unlabeled dataset and achieved a 97.01% accuracy rate which outweighs similar studies.

Semi-supervised machine learning of optical in-situ monitoring data for anomaly detection in laser powder bed fusion

ABSTRACT Laser powder bed fusion (L-PBF) is one of the most widely used metal additive manufacturing technology for fabrication of functional and structural components. However, inconsistency in quality and reliability of L-PBF products is still a significant barrier preventing it from wider adoption. Machine learning (ML) of monitoring data offers a unique solution to effectively identify possible defects and predict the quality of L-PBF products. In this work, we introduce a semi-supervised ML approach to detect anomalies that occurred in L-PBF products. We train the ML model to classify surface appearances in the reference monitoring data. We then correlate the classified appearances to post-process characteristics, e.g. surface roughness, morphology, or tensile strength. We demonstrate that the established correlation enables the determination of key appearances indicative of the quality of the printed samples including anomaly-free, lack-of-fusion and overheated. We further validate our ML approach by performing prediction on test samples having various geometries.

Walking a Tightrope: CEO Emphasis on Digital Transformation and CEO Dismissal

While CEOs’ commitment is crucial to the success of digital transformation, surveys show that CEOs of incumbent firms underemphasize the urgency and challenges of digital transformation. We suggest that career penalties may discourage CEOs from emphasizing digital transformation. We argue that this is driven by governance actors’ attribute substitution, the tendency to unconsciously substitute difficult questions with easier ones. To test these arguments, we apply a semi-supervised machine learning approach to conference calls and create a novel measure for CEOs’ emphasis on digital transformation. We find that CEOs that emphasize digital transformation have a higher probability to be dismissed if financial performance suffers. We find that this relation is stronger if actors within the firm’s governance context possess less digital and more financial expertise, and are busier. Further, we find that firms with CEOs sensitive to career concerns introduce fewer digital products and that this effect is mediated by CEOs’ emphasis on digital transformation. These results suggest that the biased evaluation of CEOs that emphasize digital transformation may impede the transformation of incumbent firms.

Who gets a say in this? Speaking security on social media

Can social media revert the top-down dynamics of securitization? Limited by the notion that ‘security is only articulated in an institutional voice by the elites’, the role of non-elite actors has remained understudied. Only recently has it been proposed that lay actors can become influential security agents through their online activity. However, social media’s capacity to revert the top-down dynamics of securitization remains contended. To explore this puzzle and seeking to update the theory of securitization to the modern context of political communication, this study employs a semi-supervised machine learning approach to analyse a novel dataset of over 10 million Twitter messages by five elite and non-elite actor groups discussing the Amazon rainforest fires in 2019. Finally, the study uses vector autoregression (VAR) models to explore who leads and who echoes the securitization process. The results show that both elite and lay actors behave as security agents and demonstrate the methodological contribution offered by the text-as-data approach developed in this analysis.