Ranking Algorithm Research Articles

Keywords, citations and ‘algorithm magic’: exploring assumptions about ranking in academic literature searches online

ABSTRACT The accessibility of academic literature has improved considerably because of the internet, with a range of platforms providing access online. It is now common for academic literature databases to use ranking algorithms to sort search results by ‘relevance’. However, it is often unclear how relevance is defined, and it varies across different platforms. This lack of transparency can potentially introduce bias, and impact the rigour of literature reviews. While there is a lack of clarity on the technical features of algorithms, online academic literature databases are now used extensively. There is a critical question of how those using the platforms perceive ranking to function in this context, and how they adapt their information-seeking behaviour. In this paper we present findings from a mixed-methods study, involving an online survey and in-depth interviews with academics, to understand their beliefs and assumptions about relevance ranking algorithms and their implications for academic practice.

Node ranking algorithm using Graph Convolutional Networks and mini-batch training

This paper presents a novel algorithm for ranking nodes in graph-structured data using Graph Convolutional Networks (GCNs) combined with mini-batch training. The proposed method integrates local and global structural information, enabling a comprehensive understanding of node importance within complex networks. By employing a multi-layer GCN architecture with residual connections and dropout regularization, our approach captures intricate graph patterns while mitigating common issues such as vanishing gradients and overfitting. The node importance scores are computed using a Multi-Layer Perceptron (MLP), with the entire model trained using Mean Squared Error (MSE) loss optimized via the Adam algorithm. We demonstrate the scalability and effectiveness of our method through extensive experiments on various benchmark datasets, showcasing its superior performance in node ranking tasks compared to existing approaches.

A Bayesian deep recommender system for uncertainty-aware online physician recommendation

Online physician recommender systems alleviate information overload by automatically recommending the best-fit physicians to patients. In contrast to general recommendations, physicians with greater uncertainty (i.e., greater variance in patients’ feedback) may not be preferred as this could affect patients’ treatment. However, most existing recommender systems don't consider uncertainty, reducing systems’ reliability and patients’ readiness to trust. To address this concern, this study leverages Bayesian theory and develops an uncertainty-aware online physician recommender system, including a Bayesian deep collaborative filtering (BDCF) model and a novel uncertainty-aware ranking algorithm. Experiments on real-world data demonstrate the superiority of BDCF and the ranking algorithm.

Music Similarity Ranking: Case Study Fairphonic

Fairphonic Pte Ltd is a technology company operating in the music sector. Fairphonic provides services to detect content on social media that is suspected of copyright infringement. Fairphonic utilizes audio features for the detection process. The current algorithm used by Fairphonic requires pairwise comparison, and the content to be compared is collected through a scraping process immediately after the process is run. Fairphonic has hundreds of thousands of music data in their database. Fairphonic desires a more scalable algorithm to compare an input music piece with the entire Fairphonic music catalog. This research uses features such as Harmonic Pitch Class Profile (HPCP), Chroma, and Rhythm Pattern. The study compares previously researched algorithms, namely binary similarity matrix, Euclidean distance, and similarity matrix profile. The results show that the combination of HPCP with the binary similarity matrix yields the highest Mean Average Precision of 0.989. Speed testing by performing comparisons 10 times shows that the combination of Chroma and the similarity matrix profile is 72% faster compared to the combination of HPCP with the binary similarity matrix. The author recommends the Chroma and similarity matrix profile algorithm for music similarity ranking due to its faster process.

Analysis of multi-criteria recommendation system based on fuzzy algorithm

There is a gap in defining the multi-criteria decision-making issues and with recommendation techniques and theories that can help develop the modulation coefficient recommenders. The main objective of this research is to identify an in-depth examination of the category of multiple variables recommendation systems. The methodology that is used in the current study is fuzzy multi-critical decision-making to enhance the precision and appropriateness of the recommendations provided to users, and make recommendations by representing an individual's performance for the product as an ordered collection of rankings in addition to different parameters. The techniques used to make forecasts and produce recommendations using multi-criteria rankings are reviewed. In addition, we propose the multiple-criteria ranking algorithms. Experimental evaluations demonstrated that our proposed algorithms can solve the multi-criteria issues. Furthermore, the research considers unresolved problems and upcoming difficulties for the category of recommendations for multiple variables ratings.

A data-centric machine learning approach to improve prediction of glioma grades using low-imbalance TCGA data

Accurate prediction and grading of gliomas play a crucial role in evaluating brain tumor progression, assessing overall prognosis, and treatment planning. In addition to neuroimaging techniques, identifying molecular biomarkers that can guide the diagnosis, prognosis and prediction of the response to therapy has aroused the interest of researchers in their use together with machine learning and deep learning models. Most of the research in this field has been model-centric, meaning it has been based on finding better performing algorithms. However, in practice, improving data quality can result in a better model. This study investigates a data-centric machine learning approach to determine their potential benefits in predicting glioma grades. We report six performance metrics to provide a complete picture of model performance. Experimental results indicate that standardization and oversizing the minority class increase the prediction performance of four popular machine learning models and two classifier ensembles applied on a low-imbalanced data set consisting of clinical factors and molecular biomarkers. The experiments also show that the two classifier ensembles significantly outperform three of the four standard prediction models. Furthermore, we conduct a comprehensive descriptive analysis of the glioma data set to identify relevant statistical characteristics and discover the most informative attributes using four feature ranking algorithms.

Extrapolation is not the same as interpolation

We propose a new machine learning formulation designed specifically for extrapolation. The textbook way to apply machine learning to drug design is to learn a univariate function that when a drug (structure) is input, the function outputs a real number (the activity): f(drug) →\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$\\rightarrow$$\\end{document} activity. However, experience in real-world drug design suggests that this formulation of the drug design problem is not quite correct. Specifically, what one is really interested in is extrapolation: predicting the activity of new drugs with higher activity than any existing ones. Our new formulation for extrapolation is based on learning a bivariate function that predicts the difference in activities of two drugs F(drug1, drug2) →\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$\\rightarrow$$\\end{document} difference in activity, followed by the use of ranking algorithms. This formulation is general and agnostic, suitable for finding samples with target values beyond the target value range of the training set. We applied the formulation to work with support vector machines , random forests , and Gradient Boosting Machines . We compared the formulation with standard regression on thousands of drug design datasets, gene expression datasets and material property datasets. The test set extrapolation metric was the identification of examples with greater values than the training set, and top-performing examples (within the top 10% of the whole dataset). On this metric our pairwise formulation vastly outperformed standard regression. Its proposed variations also showed a consistent outperformance. Its application in the stock selection problem further confirmed the advantage of this pairwise formulation.

Machine learning and experimental analyses identified miRNA expression models associated with metastatic osteosarcoma

Osteosarcoma (OS), as the most common primary bone cancer, has a high invasiveness and metastatic potential, therefore, it has a poor prognosis. This study identified early diagnostic biomarkers using miRNA expression profiles associated with osteosarcoma metastasis. In the first step, we used RNA-seq and online microarray data from osteosarcoma tissues and cell lines to identify differentially expressed miRNAs. Then, using seven feature selection algorithms for ranking, the first-ranked miRNAs were selected as input for five machine learning systems. Using network analysis and machine learning algorithms, we developed new diagnostic models that successfully differentiated metastatic osteosarcoma from non-metastatic samples based on newly discovered miRNA signatures. The results showed that miR-34c-3p and miR-154-3p act as the most promising models in the diagnosis of metastatic osteosarcoma. Validation for this model by RT-qPCR in benign tissue and osteosarcoma biopsies confirmed the lower expression of miR-34c-3p and miR-154-3p in OS samples. In addition, a direct correlation between miR-34c-3p expression, miR-154-3p expression and tumor grade was discovered. The combined values of miR-34c-3p and miR-154-3p showed 90 % diagnostic power (AUC = 0.90) for osteosarcoma samples and 85 % (AUC = 0.85) for metastatic osteosarcoma. Adhesion junction and focal adhesion pathways, as well as epithelial-to-mesenchymal transition (EMT) GO terms, were identified as the most significant KEGG and GO terms for the top miRNAs. The findings of this study highlight the potential use of novel miRNA expression signatures for early detection of metastatic osteosarcoma. These findings may help in determining therapeutic approaches with a quantitative and faster method of metastasis detection and also be used in the development of targeted molecular therapy for this aggressive cancer. Further research is needed to confirm the clinical utility of miR-34c-3p and miR-154-3p as diagnostic biomarkers for metastatic osteosarcoma.

Providing bank branch ranking algorithm with fuzzy data, using a combination of two methods DEA and MCDM

Providing bank branch ranking algorithm with fuzzy data, using a combination of two methods DEA and MCDM

Effective drug-target affinity prediction via generative active learning

Drug-target affinity (DTA) prediction is a critical early-stage task in drug discovery. Recently, deep learning has demonstrated remarkable efficacy in DTA prediction. However, acquiring experimentally verified data for target proteins proves to be a time-consuming, labor-intensive, and costly endeavor. In this study, we introduce an innovative generative active learning method for DTA prediction, referred to as GAL-DTA. GAL-DTA comprises two modules, data augmentation and generator fine-tuning. In the data augmentation module, the algorithm uses an optimized generator to produce informative and diverse molecules, thereby enhancing training of the predictor. The generator fine-tuning module introduces Fisher's informativeness and molecule diversity as objectives and employs the Pareto ranking algorithm to compute rewards. The final generator is fine-tuned using the policy-gradient method. GAL-DTA performs data augmentation by directly generating diverse and informative data, effectively reducing annotation costs while preserving model performance. Extensive experiments on independent test sets involving four target proteins consistently demonstrated that GAL-DTA achieves superior performance, resulting in an average reduction of 8.402% in mean squared error.

Improved theoretical guarantee for rank aggregation via spectral method

Abstract Given pairwise comparisons between multiple items, how to rank them so that the ranking matches the observations? This problem, known as rank aggregation, has found many applications in sports, recommendation systems and other web applications. We focus on the ranking problem under the Erdös–Rényi outliers model: only a subset of pairwise comparisons is observed, being either clean or corrupted copies of the true score differences. We investigate the spectral ranking algorithms that are based on unnormalized and normalized data matrices. The key is to understand their performance in recovering the underlying scores of each item from the observed data. This reduces to deriving an entry-wise perturbation error bound between the top eigenvectors of the unnormalized/normalized data matrix and its population counterpart. By using the leave-one-out technique, we provide a sharper $\ell _{\infty }$-norm perturbation bound of the eigenvectors and derive an error bound on the maximum displacement for each item, with only $O(n\log n)$ samples. In addition, we also derive the sample complexity to perform top-$K$ ranking under mild assumptions. Our theoretical analysis improves upon the state-of-the-art results in terms of sample complexity, and our numerical experiments confirm these theoretical findings.

A Evolution and Impact of Web Search Engines: A Comprehensive Review

Web search engines have become indispensable tools in our daily lives, revolutionizing the way we access and navigate the vast expanse of information available on the internet. This research paper provides an in-depth analysis of the evolution, functioning, and impact of web search engines. Beginning with a historical overview, it traces the development of search engines from their inception to the present day. The paper explores the underlying technologies that power search engines, including crawling, indexing, and ranking algorithms. Furthermore, it examines the socio-economic implications of search engines, discussing their influence on information retrieval, online advertising, user behavior, and societal dynamics. Finally, the paper discusses current trends and future prospects in the field of web search engines.

A novel ranking approach for identifying crucial spreaders in complex networks based on Tanimoto Correlation

The identification and rank of crucial spreaders aim to survey the diffusion capability, which has a significant impact on controlling information spread in networks. However, most of the findings neglect the fused information of the structural hole and neighbors within two orders. In addition, they fail to focus on the effect brought by the association strength between a pair of different neighborhood sets. To solve these issues, in this work, we design a novel semi-local ranking algorithm for achieving better recognition effects, referred to as the Extended Tanimoto Correlation (ETC). We first construct a topological structure of neighborhood sets Cluster, and define HierarchyVectorofCluster based on a novel layered approach IKsD. Considering the structural hole, Tanimoto Correlation (TC) is presented to contribute to measuring the association strength between different clusters. Additionally, we explicitly extend the Tanimoto Correlation metric on semi-local information to calculate each spreader’s influence score. Simulation experiments employ capability indicators and correctness metrics to apply in various networks. Results demonstrate that our method outperforms state-of-the-art algorithms in identifying and ranking effects for good spreading influence.

AC–DC security-constrained optimal power flow for the Australian National Electricity Market

This paper presents a new methodology to solve the integrated ac–dc security-constrained optimal power flow (SCOPF) problem for large-scale power systems, with a focus on the Australian National Electricity Market (NEM). The proposed SCOPF is based on a decomposition scheme that significantly reduces the dimensionality of the problem and thus improves the computational performance. It further supports a contingency ranking, evaluation, and non-dominated filtering algorithm, cyclic selection of filtered contingencies, careful choice of slack variables, and a tailored warm start to the nonlinear problem. The methodology is validated against the two-stage mathematical programming model that uses an explicit nonlinear model of both the base case and the contingency cases, at a slightly higher cost but significantly faster execution. Moreover, the NEM power system, is studied addressing different scenarios including load variations, renewable energy zone integration, and additional reinforcements like point-to-point and meshed HVDC. The proposed approaches provides preventive solutions in less than 5 min thus affirming the model’s effectiveness in real-time applications and planning studies.

Optimizing pool mining performance: A VIKOR‐based model for identifying reputed miners in blockchain networks

AbstractBlockchain networks continue to gain attraction in cutting‐edge applications and mining within these networks has become increasingly popular. To get rewards, miners solve cryptographic puzzles and add new blocks to blockchain networks using the proof‐of‐work (PoW) consensus mechanism. Numerous miners opt to participate in mining pools due to the challenges of solo mining. However, selecting reputed miners for pool mining poses a significant challenge, given the decentralized nature of the blockchain system. This paper addresses this challenge by introducing a new ranking model that evaluates miners' performance and reputation through trust scores. It provides a method for optimizing pool mining performance by identifying highly reputed miners within mining pools, enhancing overall pool profitability. This endeavor necessitates the development of ranking algorithms tailored to the unique dynamics of mining pools. The research offers a meticulously designed ranking model that identifies reputed miners. We extensively evaluate the proposed model using the hyperledger blockchain framework, guaranteeing strong performance across vital metrics like block authorization time, Processing time, block creation time, validation time, and confirmation time.

RefAI: a GPT-powered retrieval-augmented generative tool for biomedical literature recommendation and summarization.

Precise literature recommendation and summarization are crucial for biomedical professionals. While the latest iteration of generative pretrained transformer (GPT) incorporates 2 distinct modes-real-time search and pretrained model utilization-it encounters challenges in dealing with these tasks. Specifically, the real-time search can pinpoint some relevant articles but occasionally provides fabricated papers, whereas the pretrained model excels in generating well-structured summaries but struggles to cite specific sources. In response, this study introduces RefAI, an innovative retrieval-augmented generative tool designed to synergize the strengths of large language models (LLMs) while overcoming their limitations. RefAI utilized PubMed for systematic literature retrieval, employed a novel multivariable algorithm for article recommendation, and leveraged GPT-4 turbo for summarization. Ten queries under 2 prevalent topics ("cancer immunotherapy and target therapy" and "LLMs in medicine") were chosen as use cases and 3 established counterparts (ChatGPT-4, ScholarAI, and Gemini) as our baselines. The evaluation was conducted by 10 domain experts through standard statistical analyses for performance comparison. The overall performance of RefAI surpassed that of the baselines across 5 evaluated dimensions-relevance and quality for literature recommendation, accuracy, comprehensiveness, and reference integration for summarization, with the majority exhibiting statistically significant improvements (P-values <.05). RefAI demonstrated substantial improvements in literature recommendation and summarization over existing tools, addressing issues like fabricated papers, metadata inaccuracies, restricted recommendations, and poor reference integration. By augmenting LLM with external resources and a novel ranking algorithm, RefAI is uniquely capable of recommending high-quality literature and generating well-structured summaries, holding the potential to meet the critical needs of biomedical professionals in navigating and synthesizing vast amounts of scientific literature.

A Comparative Analysis of Machine Learning Algorithms in Predicting the Performance of a Combined Radiant Floor and Fan Coil Cooling System

Machine learning algorithms have proven to be practical in a wide range of applications. Many studies have been conducted on the operational energy consumption and thermal comfort of radiant floor systems. This paper conducts a case study in a self-designed experimental setup that combines radiant floor and fan coil cooling (RFCFC) and develops a data monitoring system as a source of historical operational data. Seven machine learning algorithms (extreme learning machine (ELM), convolutional neural network (CNN), genetic algorithm-back propagation (GA-BP), radial basis function (RBF), random forest (RF), support vector machine (SVM), and long short-term memory (LSTM)) were employed to predict the behavior of the RFCFC system. Corresponding prediction models were then developed to evaluate operative temperature (Top) and energy consumption (Eh). The performance of the model was evaluated using five error metrics. The obtained results showed that the RF model had very high performance in predicting Top and Eh, with high correlation coefficients (&gt;0.9915) and low error metrics. Compared with other models, it also demonstrated high accuracy in Eh prediction, yielding maximum reductions of 68.1, 82.4, and 43.2% in the mean absolute percentage error (MAPE), mean squared error (MSE), and mean absolute error (MAE), respectively. A sensitivity ranking algorithm analysis was also conducted. The obtained results demonstrated the importance of adjusting parameters, such as the radiant floor supply water temperature, to enhance the indoor comfort. This study provides a novel and effective method for evaluating the energy efficiency and thermal comfort of radiant cooling systems. It also provides insights for optimizing the efficiency and thermal comfort of RFCFC systems, and lays a theoretical foundation for future studies integrating machine learning algorithms in this field.

Effective, Efficient, and Robust Learning Algorithms for Ranking and Classification

Over the past decade, machine learning has gained significant traction and is now deployed across diverse domains, including information systems, finance, healthcare, cybersecurity, autonomous driving, and more. As machine learning finds applications in various sensitive scenarios, the demand for models that exhibit accuracy and robustness during the operational phase has grown exponentially. One crucial factor that profoundly shapes the quality of machine learning models revolves around the training data they rely upon and the input data encountered at the operational phase. Therefore, the development of data-aware algorithms is of paramount importance in achieving high-quality machine-learning models. This thesis contributes to this overarching objective by delving into the development of data-aware algorithms, emphasizing the importance of this awareness during both the training and operational phases of machine learning models. The research presented in this thesis focuses on two primary domains. The first domain is information retrieval, with a particular emphasis on enhancing both the efficiency of learning-to-rank learning algorithms and the effectiveness of the learned models in solving ranking tasks. The thesis includes three works in this domain: Marcuzzi et al. [2022] provides a novel algorithm to detect and remove consistent-outliers documents from the training data. In Marcuzzi et al. [2023], we designed a new learning algorithm that handles the problem of gradient incoherencies affecting LambdaRank-based algorithms. Finally, in Lucchese et al. [2023], we designed a new sampling function for the Selective Gradient Boosting algorithm to exploit the most useful low-ranked non-relevant document. The second domain is adversarial machine learning, which focuses on increasing the robustness of binary classifiers against adversarial inputs encountered at the operational phase. Furthermore, the research in this domain focuses on providing certifiable models to efficiently assess robustness against adversarial machine learning attacks. In this regard, in Calzavara et al. [2021], we designed a novel robust learning algorithm to train ensembles of decision trees robust to evasion attacks along with its polynomial robustness-certification algorithm designed to compute a robustness lower bound. Finally, in Calzavara et al. [2022], we provided a new evaluation metric named Resilience to better access the security of machine learning models. Awarded by : Università Ca' Foscari di Venezia, Venice, Italy on 19 April 2024. Supervised by : Claudio Lucchese. Available at : https://federicomarcuzzi.github.io/resources/thesis_phd.pdf.

EASESUM: an online abstractive and extractive text summarizer using deep learning technique

&lt;div align="center"&gt;&lt;table width="590" border="1" cellspacing="0" cellpadding="0"&gt;&lt;tbody&gt;&lt;tr&gt;&lt;td valign="top" width="387"&gt;&lt;p class="CM12"&gt;Large volumes of information are generated daily, making it challenging to manage such information. This is due to redundancy and the type of data available, most of which needs to be more structured and increases the amount of search time. Text summarization systems are considered a real solution to this vast amount of data because they are used for document compression and reduction. Text summarization keeps the relevant information and eliminates the text's non-relevant parts. This study uses two types of summarizers: Extractive Text summarizers and Abstractive text summarizers. The Text Rank Algorithm was used to implement the Extractive summarizer, while Bi-directional Recurrent Neural Network (RNN) was used to implement the Abstractive text summarizer. To improve the quality of summaries produced, word embedding was also used. For the evaluation of the summarizers, the ROUGE evaluation system was used. ROUGE contrasts summaries created by hand versus those created automatically. ROUGE examination of the produced summary revealed the superiority of human-produced summaries over those generated automatically. For this paper, a summarizer was implemented as a Web Application. The average ROUGE recall score ranging from 30.00 to 60.00 for abstractive summarizer and 0.75 to 0.82 for extractive text showed an encouraging result.&lt;/p&gt;&lt;/td&gt;&lt;/tr&gt;&lt;/tbody&gt;&lt;/table&gt;&lt;/div&gt;

Banzhaf Values for Facts in Query Answering

Quantifying the contribution of database facts to query answers has been studied as means of explanation. The Banzhaf value, originally developed in Game Theory, is a natural measure of fact contribution, yet its efficient computation for select-project-join-union queries is challenging. In this paper, we introduce three algorithms to compute the Banzhaf value of database facts: an exact algorithm, an anytime deterministic approximation algorithm with relative error guarantees, and an algorithm for ranking and top-k. They have three key building blocks: compilation of query lineage into an equivalent function that allows efficient Banzhaf value computation; dynamic programming computation of the Banzhaf values of variables in a Boolean function using the Banzhaf values for constituent functions; and a mechanism to compute efficiently lower and upper bounds on Banzhaf values for any positive DNF function. We complement the algorithms with a dichotomy for the Banzhaf-based ranking problem: given two facts, deciding whether the Banzhaf value of one is greater than of the other is tractable for hierarchical queries and intractable for non-hierarchical queries. We show experimentally that our algorithms significantly outperform exact and approximate algorithms from prior work, most times up to two orders of magnitude. Our algorithms can also cover challenging problem instances that are beyond reach for prior work.