Hochschulschriftenserver der TH Wildau
Not a member yet
1283 research outputs found
Sort by
A Comparison of Commercial Sentiment Analysis Services
No full textEmpirical insights into promising commercial sentiment analysis solutions that go beyond the claims of their vendors are rare. Moreover, due to the constant evolution in the field, previous studies are far from reflecting the current situation. The goal of this article is to evaluate and compare current solutions using two experimental studies. In the first part of the study, based on tweets about airline service quality, we test the solutions of six vendors with different market power, such as Amazon, Google, IBM, Microsoft, Lexalytics, and MeaningCloud, and report their measures of accuracy, precision, recall, (macro)F1, time performance, and service level agreements (SLA). Furthermore, we compare two of the services in depth with multiple data sets and over time. The services tested here are Google Cloud Natural Language API and MeaningCloud Sentiment Analysis API. For evaluating the results over time, we use the same data set as in November 2020. In addition, further topic-specific and general Twitter data sets are used. The experiments show that the IBM Watson NLU and Google Cloud Natural Language API solutions may be preferred when negative text detection is the primary concern. When tested in July 2022, the Google Cloud Natural Language API was still the clear winner compared to the MeaningCloud Sentiment Analysis API, but only on the airline service quality data set; on the other data sets, both services provided specific benefits and drawbacks. Furthermore, we detected changes in the sentiment classification over time with both services. Our results motivate that an independent, critical, and longitudinal experimental analysis of sentiment analysis services can provide interesting insights into their overall reliability and particular classification accuracy beyond marketing claims to critically compare solutions based on real data and analyze potential weaknesses and margins of error before making an investment
Evaluation of fabrication methods for Fabry-Perot polymer film ultrasound sensors
No full textFabry-Pèrot (FP) interferometer sensors enable highly sensitive backward mode acoustic detection in Photoacoustic (PA) imaging. They are transparent to the excitation wavelength, can be placed directly next to the PA source, and offer a broadband frequency response and high acoustic sensitivity. PA tomography using parallelized detection requires high spatial uniformity of the optical and acoustic properties, which can be hampered by contaminations during fabrication that lead to the formation of inhomogeneities and artefacts. The quality and homogeneity of the dielectric and polymer layers have a direct effect on the maximum optical phase sensitivity, and hence acoustic sensitivity. In this study, cross-sectional images of FP sensors were obtained using focused ion beam milling and ultramicrotomy followed by Scanning Electron Microscopy (SEM) and Transmission Electron Microscopy (TEM) to evaluate different fabrication methods
Using ChatGPT in Education: Human Reflection on ChatGPT’s Self-Reflection
No full textChatGPT is a fascinating AI text generator tool. It is a language model developed by OpenAI, a research and deployment company with the mission, according to OpenAI’s website: “to ensure that artificial general intelligence benefits all of humanity”. ChatGPT is able to generate human-like texts. But how does it work? What about the quality of the texts it provides? And is it capable of being self-reflective? Information sources must be efficient, effective and reliable in education, in order to enhance students’ learning process. For this reason, we started a dialogue with ChatGPT-3 while using, among others, a SWOT analysis it generated about its own functioning in an educational setting. This enabled us, as human authors, to analyze the extent to which this AI system is able to practice self-reflection. Finally, the paper sketches implications for education and future research
Understanding Website Privacy Policies—A Longitudinal Analysis Using Natural Language Processing
No full textPrivacy policies are the main method for informing Internet users of how their data are collected and shared. This study aims to analyze the deficiencies of privacy policies in terms of readability, vague statements, and the use of pacifying phrases concerning privacy. This represents the undertaking of a step forward in the literature on this topic through a comprehensive analysis encompassing both time and website coverage. It characterizes trends across website categories, top-level domains, and popularity ranks. Furthermore, studying the development in the context of the General Data Protection Regulation (GDPR) offers insights into the impact of regulations on policy comprehensibility. The findings reveal a concerning trend: privacy policies have grown longer and more ambiguous, making it challenging for users to comprehend them. Notably, there is an increased proportion of vague statements, while clear statements have seen a decrease. Despite this, the study highlights a steady rise in the inclusion of reassuring statements aimed at alleviating readers’ privacy concerns
A toolbox of machine learning software to support microbiome analysis
No full textThe human microbiome has become an area of intense research due to its potential impact on human health. However, the analysis and interpretation of this data have proven to be challenging due to its complexity and high dimensionality. Machine learning (ML) algorithms can process vast amounts of data to uncover informative patterns and relationships within the data, even with limited prior knowledge. Therefore, there has been a rapid growth in the development of software specifically designed for the analysis and interpretation of microbiome data using ML techniques. These software incorporate a wide range of ML algorithms for clustering, classification, regression, or feature selection, to identify microbial patterns and relationships within the data and generate predictive models. This rapid development with a constant need for new developments and integration of new features require efforts into compile, catalog and classify these tools to create infrastructures and services with easy, transparent, and trustable standards. Here we review the state-of-the-art for ML tools applied in human microbiome studies, performed as part of the COST Action ML4Microbiome activities. This scoping review focuses on ML based software and framework resources currently available for the analysis of microbiome data in humans. The aim is to support microbiologists and biomedical scientists to go deeper into specialized resources that integrate ML techniques and facilitate future benchmarking to create standards for the analysis of microbiome data. The software resources are organized based on the type of analysis they were developed for and the ML techniques they implement. A description of each software with examples of usage is provided including comments about pitfalls and lacks in the usage of software based on ML methods in relation to microbiome data that need to be considered by developers and users. This review represents an extensive compilation to date, offering valuable insights and guidance for researchers interested in leveraging ML approaches for microbiome analysis
Electron correlation dynamics in atomic Kr excited by XUV pulses and controlled by NIR laser pulses of variable intensity
No full textWe have investigated the possibility to track and control correlation dynamics of valence electrons in krypton (Kr) initiated by the absorption of one extreme ultraviolet (XUV) photon. In this investigation, pump-probe experiments have been performed where monochromatized single high-harmonics at photon energies 29.6, 32.8, and 35.9 eV have been used as pump to populate different intermediate excited states. A temporally delayed near-infrared (NIR) pulse probes the population of various decay channels via the detection of Kr²⁺ ion yields and its transient profiles. We observe that by varying the NIR pulse intensity within a range from 0.3 x 10¹³ to 2.6 x 10¹³ W cm⁻², the shape of the Kr²+ transient profile changes significantly. We show that by varying the intensity of the NIR pulse, it is possible—(i) to control the ratio between sequential and non-sequential double ionization of Kr; (ii) to selectively probe quantum beating oscillations between Kr+* satellite states that are coherently excited within the bandwidth of the XUV pulse; and (iii) to specifically probe the relaxation dynamics of doubly excited (Kr**) decay channels. Our studies show that the contribution of different ionization and decay channels (i)–(iii) can be altered by the NIR pulse intensity, thus demonstrating an efficient way to control the ionization dynamics in rare gas atoms
Closing the green gap of photosystem I with synthetic fluorophores for enhanced photocurrent generation in photobiocathodes
No full textOne restriction for biohybrid photovoltaics is the limited conversion of green light by most natural photoactive components. The present study aims to fill the green gap of photosystem I (PSI) with covalently linked fluorophores, ATTO 590 and ATTO 532. Photobiocathodes are prepared by combining a 20 μm thick 3D indium tin oxide (ITO) structure with these constructs to enhance the photocurrent density compared to setups based on native PSI. To this end, two electron transfer mechanisms, with and without a mediator, are studied to evaluate differences in the behavior of the constructs. Wavelength-dependent measurements confirm the influence of the additional fluorophores on the photocurrent. The performance is significantly increased for all modifications compared to native PSI when cytochrome c is present as a redox-mediator. The photocurrent almost doubles from −32.5 to up to −60.9 μA cm−2. For mediator-less photobiocathodes, interestingly, drastic differences appear between the constructs made with various dyes. While the turnover frequency (TOF) is doubled to 10 e−/PSI/s for PSI-ATTO590 on the 3D ITO compared to the reference specimen, the photocurrents are slightly smaller since the PSI-ATTO590 coverage is low. In contrast, the PSI-ATTO532 construct performs exceptionally well. The TOF increases to 31 e−/PSI/s, and a photocurrent of −47.0 μA cm−2 is obtained. This current is a factor of 6 better than the reference made with native PSI in direct electron transfer mode and sets a new record for mediator-free photobioelectrodes combining 3D electrode structures and light-converting biocomponents
A Data-Driven Approach Towards the Application of Reinforcement Learning Based HVAC Control
No full textRefrigeration applications consume a significant share of total electricity demand, with a high indirect impact on global warming through greenhouse gas emissions. Modern technology can help reduce the high power consumption and optimize the cooling control. This paper presents a case study of machine-learning for controlling a commercial refrigeration system. In particular, an approach to reinforcement learning is implemented, trained and validated utilizing a model of a real chiller plant. The reinforcement-learning controller learns to operate the plant based on its interactions with the modeled environment. The validation demonstrates the functionality of the approach, saving around 7% of the energy demand of the reference control. Limitations of the approach were identified in the discretization of the real environment and further model-based simplifications and should be addressed in future research
Artificial neural network in soft HR performance management: new insights from a large organizational dataset
No full textPurpose
This study investigates whether the artificial neural network approach, when used on a large organizational soft HR performance dataset, results in a better (R2/RMSE) model compared to the linear regression. With the use of predictive modelling, a more informed base for managerial decision making within soft HR performance management is offered.
Design/methodology/approach
The study builds on a dataset (n > 43 k) stemming from an annual employee MNC survey. It covers several soft HR performance drivers and outcomes (such as engagement, satisfaction and others) that either have evidence of a dual-role nature or non-linear relationships. This study applies the framework for artificial neural network analysis in organization research (Scarborough and Somers, 2006).
Findings
The analysis reveals a substantial artificial neural network model performance (R2 > 0.75) with an excellent fit statistic (nRMSE <0.10) and all drivers have the same relative importance (RMI [0.102; 0.125]). This predictive analysis revealed that the organization has to increase six of the drivers, keep two on the same level and decrease one.
Originality/value
Up to date, this study uses the largest dataset in soft HR performance management. Additionally, the predictive results reveal that specific target values lay below the current levels to achieve optimal performance
IT-Assisted Optimisation of Fuel Consumption in Air Transport
No full textAviation continues to be an essential means of transport for passengers and cargo. In recent years, the COVID-19 pandemic led to a collapse. In 2022 Europe was back up to around 85% of 2019 levels(EASA, 2022).In 2021 van der Sman et al. predicted an recovery to 2019 levels in 2024 (van der Sman et al., 2021). However, fuel savings and emissions reduction have become increasingly important in recent years.
As part of a PhD thesis, possibilities for reducing fuel consumption by reducing the final reserve fuel were investigated. A smaller amount of tanked fuel required leads to a reduction in the transported (fuel) weight and, thus, a reduction in overall fuel consumption. This is because fuel consumption for a given route depends, among other factors, on the aircraft's weight. The more an aircraft weighs, the higher the fuel consumption. To keep fuel consumption as low as possible, carrying only the minimum weight required for the route in question is the most economical. Carrying more or even unnecessary weight increases the amount of fuel required and consumed in flight.
The overarching research aims to explore and evaluate how to reduce the fuel carried by aircraft and, thus, the total fuel required for a given flight. The main focus of this paper is on the opportunities and challenges that have arisen with introducing new fuel regulations in European aviation regulations. Operators with appropriate safety levels can apply more tailored provisions. This requires the demonstration of the safety level. This is achieved by defining specific safety performance indicators (SPIs), compliance with which is then continuously monitored and evaluated during operation. This requires the collection and evaluation of correspondingly large amounts of data. This is only possible using appropriate IT applications. Example belowshows the amount of data that accumulates during flight operations. Recording, processing and saving pose a challenge in this respect.
On the other hand, performance-based regulations allow for a more individualised implementation on and by the respective companies via the demonstration of a corresponding level of safety. Safety indicators are used for this purpose, which must be obtained and evaluated from various existing data. A large amount of data, which can only be collected and processed with the help of various IT applications, represents a challenge. However, companies can benefit from the corresponding advantages if they can cope with this.
The following is an excerpt of the requirements and possible implementation, focusing on the amount of data and the associated challenges and opportunities