Machine Learning Approach Research Articles

Attack and Anomaly Detection in IoT Sensors Using Machine Learning Approaches

Attack and Anomaly Detection in IoT Sensors Using Machine Learning Approaches

Prediction of electrohydrodynamic printing behavior using machine learning approaches

Prediction of electrohydrodynamic printing behavior using machine learning approaches

Molecular Insights into α-Synuclein Fibrillation: A Raman Spectroscopy and Machine Learning Approach.

The aggregation of α-synuclein is crucial to the development of Lewy body diseases, including Parkinson's disease and dementia with Lewy bodies. The aggregation pathway of α-synuclein typically involves a defined sequence of nucleation, elongation, and secondary nucleation, exhibiting prion-like spreading. This study employed Raman spectroscopy and machine learning analysis, alongside complementary techniques, to characterize the biomolecular changes during the fibrillation of purified recombinant wild-type α-synuclein protein. Monomeric α-synuclein was produced, purified, and subjected to a 7-day fibrillation assay to generate preformed fibrils. Stages of α-synuclein fibrillation were analyzed using Raman spectroscopy, with aggregation confirmed through negative staining transmission electron microscopy, mass spectrometry, and light scattering analyses. A machine learning pipeline incorporating principal component analysis and uniform manifold approximation and projection was used to analyze the Raman spectral data and identify significant peaks, resulting in differentiation between sample groups. Notable spectral shifts in α-synuclein were found in various stages of aggregation. Early changes (D1) included increases in α-helical structures (1303, 1330 cm-1) and β-sheet formation (1045 cm-1), with reductions in COO- and CH2 bond regions (1406, 1445 cm-1). By D4, these structural shifts persist with additional β-sheet features. At D7, a decrease in β-sheet H-bonding (1625 cm-1) and tyrosine ring breathing (830 cm-1) indicates further structural stabilization, suggesting a shift from initial helical structures to stabilized β-sheets and aggregated fibrils. Additionally, alterations in peaks related to tyrosine, alanine, proline, and glutamic acid were identified, emphasizing the role of these amino acids in intramolecular interactions during the transition from α-helical to β-sheet conformational states in α-synuclein fibrillation. This approach offers insight into α-synuclein aggregation, enhancing the understanding of its role in Lewy body disease pathophysiology and potential diagnostic relevance.

PM6-ML: The Synergy of Semiempirical Quantum Chemistry and Machine Learning Transformed into a Practical Computational Method.

Machine learning (ML) methods offer a promising route to the construction of universal molecular potentials with high accuracy and low computational cost. It is becoming evident that integrating physical principles into these models, or utilizing them in a Δ-ML scheme, significantly enhances their robustness and transferability. This paper introduces PM6-ML, a Δ-ML method that synergizes the semiempirical quantum-mechanical (SQM) method PM6 with a state-of-the-art ML potential applied as a universal correction. The method demonstrates superior performance over standalone SQM and ML approaches and covers a broader chemical space than its predecessors. It is scalable to systems with thousands of atoms, which makes it applicable to large biomolecular systems. Extensive benchmarking confirms PM6-ML's accuracy and robustness. Its practical application is facilitated by a direct interface to MOPAC. The code and parameters are available at https://github.com/Honza-R/mopac-ml.

Fast and interpretable mortality risk scores for critical care patients.

Prediction of mortality in intensive care unit (ICU) patients typically relies on black box models (that are unacceptable for use in hospitals) or hand-tuned interpretable models (that might lead to the loss in performance). We aim to bridge the gap between these 2 categories by building on modern interpretable machine learning (ML) techniques to design interpretable mortality risk scores that are as accurate as black boxes. We developed a new algorithm, GroupFasterRisk, which has several important benefits: it uses both hard and soft direct sparsity regularization, it incorporates group sparsity to allow more cohesive models, it allows for monotonicity constraint to include domain knowledge, and it produces many equally good models, which allows domain experts to choose among them. For evaluation, we leveraged the largest existing public ICU monitoring datasets (MIMIC III and eICU). Models produced by GroupFasterRisk outperformed OASIS and SAPS II scores and performed similarly to APACHE IV/IVa while using at most a third of the parameters. For patients with sepsis/septicemia, acute myocardial infarction, heart failure, and acute kidney failure, GroupFasterRisk models outperformed OASIS and SOFA. Finally, different mortality prediction ML approaches performed better based on variables selected by GroupFasterRisk as compared to OASIS variables. Group Faster Risk's models performed better than risk scores currently used in hospitals, and on par with black box ML models, while being orders of magnitude sparser. Because GroupFasterRisk produces a variety of risk scores, it allows design flexibility-the key enabler of practical model creation. Group Faster Risk is a fast, accessible, and flexible procedure that allows learning a diverse set of sparse risk scores for mortality prediction.

Predictive role of SLC1A5 in neuroblastoma prognosis and immunotherapy

BackgroundNeuroblastoma, a prevalent extracranial solid tumor in pediatric patients, demonstrates significant clinical heterogeneity, ranging from spontaneous regression to aggressive metastatic disease. Despite advances in treatment, high-risk neuroblastoma remains associated with poor survival. SLC1A5, a key glutamine transporter, plays a dual role in promoting tumor growth and immune modulation. However, its contributions to neuroblastoma biology remain largely unexplored.MethodsThis study utilized clinical neuroblastoma samples from 20 patients and 1310 cases from four public datasets to investigate SLC1A5 expression, biological function, and prognostic significance. Differential expression, Kaplan–Meier survival analysis, gene set enrichment analysis, and weighted correlation network analysis were conducted. Functional validation included qPCR, immunohistochemistry, Western blotting, and cell proliferation assays using the SLC1A5 inhibitor V-9302. A prognostic signature, SRPS, was constructed and validated using machine-learning approaches. Immune infiltration analysis was performed to evaluate the tumor immune microenvironment.ResultsSLC1A5 expression was significantly elevated in high-risk neuroblastoma and correlated with advanced stages and poor prognosis. GSEA revealed mTORC1 signaling enrichment in high SLC1A5 expression groups, validated by increased p-p70S6K levels in tumor samples and neuroblastoma cells. V-9302 treatment suppressed mTORC1 signaling and inhibited cell proliferation. Hub-genes were identified to form the SRPS model, which demonstrated superior prognostic performance compared to existing models. Immune infiltration analysis revealed a more immunosuppressive tumor microenvironment associated with high SLC1A5 expression. Additionally, SLC1A5 negatively regulated ST8SIA1, a gene crucial for GD2 biosynthesis, suggesting that SLC1A5 inhibition may enhance GD2-directed immunotherapies.ConclusionSLC1A5 plays a pivotal role in neuroblastoma by promoting tumor progression and shaping an immunosuppressive microenvironment. The SRPS model, incorporating SLC1A5-associated genes, offers robust prognostic utility. Targeting SLC1A5 through advanced drug delivery systems and combined metabolic-immunotherapeutic strategies may enhance treatment specificity and efficacy. These findings provide a foundation for novel therapeutic approaches to improve outcomes in high-risk neuroblastoma patients.

Assessing machine learning and data imputation approaches to handle the issue of data sparsity in sports forecasting

Assessing machine learning and data imputation approaches to handle the issue of data sparsity in sports forecasting

Machine Learning Approaches for Effective Credit Card Fraud Detection: Addressing Imbalance and Enhancing Accuracy

Machine Learning Approaches for Effective Credit Card Fraud Detection: Addressing Imbalance and Enhancing Accuracy

BART-Survival: A Bayesian machine learning approach to survival analyses in Python

BART-Survival: A Bayesian machine learning approach to survival analyses in Python

Re-scaling images using a SVD-based approach

There are many instances in computer science where computational operations must be performed on matrices of different sizes. In the field of machine learning, particularly when interpreting images, it is often necessary to resize and re-scale images to achieve higher resolutions. While there are various methods for this, they typically involve fitting a simple linear or cubic model to scale the image. The singular value decomposition (SVD) is a powerful tool for dimension reduction and projection. Our proposed state-of-the-art method leverages the capabilities of SVD to create a technique for re-scaling images. This method is primarily based on modifications of the eigenvectors derived from SVD. Previous work has shown that by editing only these Eigenvectors, it is possible to minimize error propagation through the images. When applied to several well-known image processing tasks, it is possible to scale an image with reduced error compared to the current state-of-the-art methods. Additionally, we show that our method can improve the results of machine learning approaches.

Machine Learning Approach for Predicting Perfluorooctanesulfonate Rejection in Efficient Nanofiltration Treatment and Removal

Machine Learning Approach for Predicting Perfluorooctanesulfonate Rejection in Efficient Nanofiltration Treatment and Removal

Improving molecular subtypes and prognosis of pancreatic cancer through multi group analysis and machine learning

BackgroundPancreatic cancer (PAC) has a complex tumor immune microenvironment, and currently, there is a lack of accurate personalized treatment. Establishing a novel consensus machine learning driven signature (CMLS) that offers a unique predictive model and possible treatment targets for this condition was the goal of this study.MethodsThis study integrated multiple omics data of PAC patients, applied ten clustering techniques and ten machine learning approaches to construct molecular subtypes for PAC, and created a new CMLS.ResultsUsing multi-omics clustering, we discovered two cancer subtypes (CSs) associated with prognosis, among which CS1 exhibited poor prognostic outcomes. Subsequently, 13 central genes were identified through screening, constituting CMLS with a significant prognostic ability. The low CMLS group had a better prognosis and was more likely to possess a “hot” tumor phenotype. The prognosis for the high CMLS group was dismal. Still, the tumor mutation burden (TMB) and tumor neoantigen burden (TNB) levels in this group of patients were higher than in the low CMLS group, which were more favorable for immune therapy response.ConclusionThis study emphasizes that CMLS provides a beneficial instrument for early prediction of patient prognosis and screening of probable patients appropriate for immunotherapy and has broad implications for clinical practice.

Evaluating the effectiveness of ensemble machine learning approaches for pore pressure prediction using petrophysical log data in carbonate reservoir

Evaluating the effectiveness of ensemble machine learning approaches for pore pressure prediction using petrophysical log data in carbonate reservoir

A Machine Learning Approach for Load Margin Assessment in Balanced Islanded Microgrids

A Machine Learning Approach for Load Margin Assessment in Balanced Islanded Microgrids

18F]PSMA-1007 PET/CT-based radiomics may help enhance the interpretation of bone focal uptakes in hormone-sensitive prostate cancer patients.

We hypothesised that applying radiomics to [18F]PSMA-1007 PET/CT images could help distinguish Unspecific Bone Uptakes (UBUs) from bone metastases in prostate cancer (PCa) patients. We compared the performance of radiomic features to human visual interpretation. We retrospectively analysed 102 hormone-sensitive PCa patients who underwent [18F]PSMA-1007 PET/CT and exhibited at least one focal bone uptake with known clinical follow-up (reference standard). Using matRadiomics, we extracted features from PET and CT images of each bone uptake and identified the best predictor model for bone metastases using a machine-learning approach to generate a radiomic score. Blinded PET readers with low (n = 2) and high (n = 2) experience rated each bone uptake as either UBU or bone metastasis. The same readers performed a second read three months later, with access to the radiomic score. Of the 178 [18F]PSMA-1007 bone uptakes, 74 (41.5%) were classified as PCa metastases by the reference standard. A radiomic model combining PET and CT features achieved an accuracy of 84.69%, though it did not surpass expert PET readers in either round. Less-experienced readers had significantly lower diagnostic accuracy at baseline (p < 0.05) but improved with the addition of radiomic scores (p < 0.05 compared to the first round). Radiomics might help to differentiate bone metastases from UBUs. While it did not exceed expert visual assessments, radiomics has the potential to enhance the diagnostic accuracy of less-experienced readers in evaluating [18F]PSMA-1007 PET/CT bone uptakes.

A compact ultrawide bandpass filter along notch characteristics with rectangular resonator through a machine learning approach

Abstract This article presents an ultrawide bandpass filter structure developed along a notch band using a small rectangular impedance resonator. The proposed filter structure consists of a coupled rectangular resonator (CRR), open stub, and composited split ring resonator (CSRR) at the bottom of the structure. In-band and out-of-band properties are improved by the CRR and open stub. The notch band is obtained by placing CSRR below the rectangular resonator. A filter with a compact size of 0.15 × 0.10 λg is obtained at a lowered cutoff frequency of 3.0 GHz, where λg is the corresponding guided wavelength. The proposed structure has been constructed on 5880 Rogers substrate with a thickness of 0.787 mm and a dielectric constant of 2.2. Additionally, equivalent lumped parameters were obtained, and a lumped equivalent circuit was created to explain how the suggested filter operated. The Electromagnetic (EM)-simulated results are in good agreement with the circuit-simulated and measured result. The various machine learning approaches such as artificial neural network, K-nearest neighbour, decision tree, random forest (RF), and extreme gradient boosting algorithms are applied to optimize the design, in which RF algorithms achieve more than 90% accuracy for predicting the S parameters of the ultrawideband filter.

Tourists Behavior Prediction through Online Reviews by Analyzing their Sentiments using Machine Learning Approach

Sentiment Analysis is that part of research area where different people opinions or sentiments are extracted in form of textual data from various websites. In this paper sentiment analysis has been described along with machine learning techniques on tourist’s reviews to see their behavior towards various tourist places, hotels and other services provided by tourism industry. Emotions in the form of tourist’s reviews extracted are interpreted and classified by preprocessing of data and further feature extraction is done through machine learning highly efficient technique called deep learning. In this paper, the proposed idea has been given to use deep learning methods like CNN, RNN and LSTM rather than using machine learning classical algorithms like SVM, Naive Bayes, KNN, RF etc. Also, comparison of various machine learning and deep learning techniques working on tourist sentiments has been done here in this paper to show that deep learning techniques analyze and classify emotions and polarity with deep layers efficiently where on the other hand classical algorithms of machine learning give results not better than deep learning techniques. In this way sentiment analysis has been done and the proposed idea of this research paper is change in the machine learning techniques or methods from classical algorithms to neural network deep learning methods which in future definitely will give better results to analyze deeply the sentiments of tourists to find out the liking and disliking of various tourist places, hotels and related tourism services that will help tourism business industry to work on the gap in existing services provided by them and system can become more efficient in future. Such improved tourism system will give benefits to tourists or users in terms of better services and undoubtedly it will help tourism industry to enhance business in future. Keywords—sentiment analysis, machine learning, deep learning, tourist reviews.

Impact of Organizational Innovativeness Practices on Performance of Commercial State-Owned Enterprises in Kenya: Insights from a Machine Learning Approach

Impact of Organizational Innovativeness Practices on Performance of Commercial State-Owned Enterprises in Kenya: Insights from a Machine Learning Approach

Machine‐learning‐based integration of tumor microenvironment features predicting immunotherapy response

ABSTRACTImmunotherapy has revolutionized cancer treatment in recent years, yet non‐responsiveness of immunotherapy remains a challenge for cancer treatment. Therefore, the prediction method for potential clinical benefits of patients from immunotherapy is urgently needed. This study aims to develop an effective clinical practice assistance tool to evaluate the potential clinical benefits and therapy responsiveness of patients undergoing immunotherapy. We developed an immunotherapy resistance score (IRS), which performed well compared with conventional immunotherapy response indicators across different immunotherapy cohorts. Tumor microenvironment (TME) analysis showed that both immune and nonimmune features collectively impact immunotherapy responsiveness. Thus, IRS was constructed based on the TME features using machine learning approaches. The clinical application potential of IRS has been demonstrated in our in‐house Harbin Medical University (HMU) cohort and an external validation cohort. Furthermore, we analyzed the correlation between IRS and pathways related to cancer therapy targets to explore the application potential of IRS in comprehensive cancer therapy. In conclusion, IRS is a robust tool for predicting patient immunotherapy prognosis, which has great potential to promote precise clinical therapy.

A Machine Learning Approach Using Topic Modeling to Identify and Assess Experiences of Patients With Colorectal Cancer: Explorative Study.

The rising number of cancer survivors and the shortage of health care professionals challenge the accessibility of cancer care. Health technologies are necessary for sustaining optimal patient journeys. To understand individuals' daily lives during their patient journey, qualitative studies are crucial. However, not all patients wish to share their stories with researchers. This study aims to identify and assess patient experiences on a large scale using a novel machine learning-supported approach, leveraging data from patient forums. Forum posts of patients with colorectal cancer (CRC) from the Cancer Survivors Network USA were used as the data source. Topic modeling, as a part of machine learning, was used to recognize the topic patterns in the posts. Researchers read the most relevant 50 posts on each topic, dividing them into "home" or "hospital" contexts. A patient community journey map, derived from patients stories, was developed to visually illustrate our findings. CRC medical doctors and a quality-of-life expert evaluated the identified topics of patient experience and the map. Based on 212,107 posts, 37 topics and 10 upper clusters were produced. Dominant clusters included "Daily activities while living with CRC" (38,782, 18.3%) and "Understanding treatment including alternatives and adjuvant therapy" (31,577, 14.9%). Topics related to the home context had more emotional content compared with the hospital context. The patient community journey map was constructed based on these findings. Our study highlighted the diverse concerns and experiences of patients with CRC. The more emotional content in home context discussions underscores the personal impact of CRC beyond clinical settings. Based on our study, we found that a machine learning-supported approach is a promising solution to analyze patients' experiences. The innovative application of patient community journey mapping provides a unique perspective into the challenges in patients' daily lives, which is essential for delivering appropriate support at the right moment.