A performance evaluation of neural network features and functions settings on the model accuracy

Bozdech, M.

doi:10.1101/2022.11.28.518263

Home/Archive/bioRxiv

bioRxiv · 10.1101/2022.11.28.518263

A performance evaluation of neural network features and functions settings on the model accuracy

Bozdech, M.

View original source ↗Export metadata ↓

Abstract

Not only in sports is a neural network the most used type of artificial intelligence. With software development, anyone can create a neural network model, but little is known about how to prepare the data and how to set up the model algorithms to their maximum performance. For these reasons, this study aims to determine whether features or function settings have a greater effect on model accuracy. An initial feature dataset (n = 18882) was obtained from publicly available sources. Each of the six different feature settings consisted of 96 models. A total of 384 models were created, in which their testing accuracy and the percentage difference between the training and testing phases were further analyzed. No statistically significant differences were found between the accuracy of the functions settings, but statistically significant differences were confirmed between the feature settings. The study found that feature settings, especially the reduction of the number of outputs, are a more important factor in increasing the model accuracy, than function settings. Although the literature focuses more on the function setting and sets feature setting is taken rather as a type of how to improve the model.

Explore related subjects

scientific communication and education Preprint

Keep this discovery

Explore connections, maps & timelines

BibTeX RIS

Bozdech, M.. 2022-11-30. A performance evaluation of neural network features and functions settings on the model accuracy. https://doi.org/10.1101/2022.11.28.518263

Cite the original work for its findings. Save a collection to share your selection of sources.

KEEP EXPLORING

Related preprints

bioRxiv2026

Evaluating Large Language Models as Tools to Navigate Researchers in Rapidly Evolving Research Landscapes: A Case Study in Cancer Drug Response Prediction

Large Language Models (LLMs) have emerged as promising tools for assisting researchers in automating and accelerating the synthesis of literature reviews. However, their reliability is a significant concern due to issues like factual inaccuracies and hallucinations. The key question is whether LLMs can reliably provide comprehensive, up-to-date overviews and analyses. This study evaluates the performance of three leading LLMs (OpenAI's ChatGPT, Google's Gemini, and DeepSeek) on the complex task of generating a comprehensive survey paper on deep learning for cancer Drug Response Prediction (DRP). By testing both standard and Deep Research (DR) / Deep Think (DT) modes of LLMs with prompts of varying detail, this paper assesses key academic dimensions, including reference management, content quality, and analytical depth. Key findings reveal that while DR modes of LLMs significantly improve reliability by eliminating hallucinations, performance variations exist across models and prompts. A trade-off between reference quantity and integration quality was observed, and even the best-performing models lacked the analytical depth of human experts, often requiring extensive human supervision. The study concludes that LLMs currently serve as powerful assistive tools but still cannot replace the critical validation and synthesis provided by human researchers. Choosing the best LLM to use depends on the task in hand, while several strategies can be implemented to improve the produced output.

scientific communication and education↗

bioRxiv2024

Enabling preprint discovery, evaluation, and analysis with Europe PMC

Preprints provide an indispensable tool for rapid and open communication of early research findings. Preprints can also be revised and improved based on scientific commentary uncoupled from journal-organised peer review. The uptake of preprints in the life sciences has increased significantly in recent years, especially during the COVID-19 pandemic, when immediate access to research findings became crucial to address the global health emergency. With ongoing expansion of new preprint servers, improving discoverability of preprints is a necessary step to facilitate wider sharing of the science reported in preprints. To address the challenges of preprint visibility and reuse, Europe PMC, an open database of life science literature, began indexing preprint abstracts and metadata from several platforms in July 2018. Since then, Europe PMC has continued to increase coverage through addition of new servers, and expanded its preprint initiative to include the full text of preprints related to COVID-19 in July 2020 and then the full text of preprints supported by the Europe PMC funder consortium in April 2022. The preprint collection can be searched via the website and programmatically, with abstracts and the open access full text of COVID-19 and Europe PMC funder preprint subsets available for bulk download in a standard machine-readable JATS XML format. This enables automated information extraction for large-scale analyses of the preprint corpus, accelerating scientific research of the preprint literature itself. This publication describes steps taken to build trust, improve discoverability, and support reuse of life science preprints in Europe PMC. Here we discuss the benefits of indexing preprints alongside peer-reviewed publications, and challenges associated with this process.

scientific communication and education↗

bioRxiv2024

Bacttle: a microbiology educational board game for lay public and schools

1.Inspired by the positive impact of serious games on science understanding and motivated by personal interests in scienti[fi]c outreach, we developed Bacttle, an easy-to-play microbiology board game with adaptive difficulty, targeting any player from 7 years old onward. Bacttle addresses both the lay public and teachers for use in classrooms as a way of introducing microbiology concepts. The layout of the game and its mechanism are the result of multiple rounds of trial, feedback and re-design. The [fi]nal version consists of a deck of cards, a 3D-printed board and tokens (with a paper-based alternative), with all digital content open source. Players in Bacttle take on the character of a bacterial species. The aim for each species is to proliferate under the environmental conditions of the board and the interactions with the board and with other players, which vary as the play evolves. Players start with a given number of lives that will increase or decrease based on the traits they play for different environmental scenarios. Such bacterial traits come in the form of cards that can be deployed strategically. In order to assess the impact of the game on microbiological knowledge, we scored differences in the understanding of general concepts before and after playing the game. We assessed a total of 169 visitors at two different university open day science fairs. Players were asked to [fi]ll a brief survey before and after the game with questions targeting conceptual advances. Results show that Bacttle increases general microbiology knowledge on players as young as 5 years old, and with the highest impact on those who have no a priori microbiology comprehension.

scientific communication and education↗