Data Mining Algorithms Research Articles

Application of an improved LightGBM hybrid integration model combining gradient harmonization and Jacobian regularization for breast cancer diagnosis

Cancer, as a shocking disease, is one of the most common malignant tumors among women, posing a huge threat to the physical health and safety of women worldwide. With the continuous development of science and technology, more and more high and new technologies are involved in the diagnosis and prediction of breast cancer. In recent years, intelligent medical assistants supported by data mining and machine learning algorithms have provided necessary support for doctors’ diagnosis. This study proposes an improved LightGBM hybrid integration model. Introducing gradient harmonic loss and cross entropy loss to enhance the model’s attention to minority classes in the dataset and alleviate the impact of data imbalance on diagnostic results. Designing whale optimization algorithm to improve LightGBM to achieve iterative optimization of hyperparameters, and enhance the overall performance of the model. Proposing Jacobian regularization method to denoise LightGBM to solve the problem of model sensitivity to noise. Developing the LightGBM hybrid integration model to ensure the accuracy and stability of model diagnosis on diverse and imbalanced datasets. The effectiveness of the proposed method has been comprehensively compared and verified through the dataset in the UCI machine learning repository, and the results show that the proposed method has achieved good diagnostic performance in all indicators. The hybrid integration model proposed in this paper can provide effective auxiliary support for doctors to diagnose breast cancer.

Evaluation of the quality of the electrical energy supplied by a grid-connected solar photovoltaic plant using machine learning and data mining algorithms

La presencia de plantas solares fotovoltaicas para la producción de electricidad implica la disminución del uso de combustibles fósiles, y la reducción de las emisiones contaminantes. La disponibilidad de la energía solar depende de las condiciones climáticas, por lo que los parámetros de la energía eléctrica a entregar podrían verse afectados. El objetivo de esta investigación es mostrar una metodología basada en la ciencia de datos para la evaluación de la calidad de la energía de plantas solares fotovoltaicas conectadas a la red, considerando los estándares vigentes. Se aplica a una planta de 260 kWp del Instituto Nacional de Estándares y Tecnología de los Estados Unidos. Los parámetros utilizados son la distorsión armónica total, fluctuaciones y desbalance de voltaje, frecuencia eléctrica, y factor de potencia. Casi el 100% de los registros cumplen con los límites establecidos por los estándares para los parámetros, a excepción del factor de potencia, con un 63,56%. Del modelo de clasificación del factor de potencia se obtuvo que la potencias aparente y activa, y la frecuencia, son las variables más importantes. Del algoritmo de descubrimiento de subgrupos, se obtuvo que la irradiancia solar aparece en el 40% de los subgrupos, y la frecuencia en el 50%.

Use of data mining algorithms in prediction of eggshell thickness from egg quality traits of Potchefstroom Koekoek layers

Egg quality is affected by lot of factors. Study was conducted to compare performance of data mining algorithms; Classification and regression tree (CART), Chi-square automatic interaction detection (CHAID), Exhaustive chi-square automatic interaction detection (Ex-CHAID) and Multivariate adaptive regression spline (MARS) in prediction of Potchefstroom Koekoek’s eggshell thickness from egg quality traits. 350 eggs were collected at 31st to 39th week to examine the egg quality traits. MARS with R2(0.86) revealed yolk ratio, shell weight, egg shape index, yolk ratio, shell ratio, albumen weight and albumen ratio as explanatory variables predicting eggshell thickness. CART with R2 (0.37), yolk/albumen ratio was noted to be influential predictor of eggshell thickness. CHAID and Ex-CHAID (R2= 0.35) discovered egg weight as the best predictor of eggshell thickness. MARS with R2(0.86) revealed yolk ratio, shell weight, egg shape index, yolk ratio, shell ratio, albumen weight and albumen ratio as explanatory variables predicting eggshell thickness. MARS had high r (0.925), R2 (0.856) and lower RMSE (0.129) and AIC (-975.331) compared to CHAID, Ex-CHAID and CART leading MARS to be the best data mining algorithm when predicting the eggshell thickness using egg quality traits.

Presenting a model to reduce students' academic drop by using analytical comparison of machine learning algorithms in data mining (case study of Shahed University)

Presenting a model to reduce students' academic drop by using analytical comparison of machine learning algorithms in data mining (case study of Shahed University)

Topic Modeling of a Text Document Using Processed Document Term Matrix (PDTM)

Topic modeling is a method of determining the topic of a text document by analyzing the semantics and syntax of the latter. When analyzing text, the method determines the internal structure of a document or a set of documents and uses this information to classify or group similar words by topic. It also helps to identify the main trends of interests or information in a text document. For example, many people are interested in online shopping, politics, sports, economics, society, and etc. There are various online and offline data mining methods and algorithms used to determine the topic of a text. Most of them use a certain mechanism based on the semantic characteristics of the language and the subject of the text. In this study, the main idea is to develop a methodology that can be effectively used for topic modeling of a text in different languages. At first,the model preprocesses a text, which includes its tokenization, deletion of STOPWORDS and its lemmatization. Text preprocessing and filtering of inappropriate text elements reduces the size of the text and improves its classification performance. The algorithm also assumes the presence of ‘n’ topics in a text document and, based on this assumption, generates the processed document term matrix (PDTM) for a text document. The Processed Document Term Matrix (PDTM) is a two-dimensional matrix that assigns a specific numerical value to each word in the text based on the frequency of its occurrence in the document, and then correlates this word with each topic assumed earlier. The processed document terms (PDTM) are generated to store tokenized words. The proposed model and its results are described in detail in the methodology and discussion sections of this article.

Real-world study of isavuconazole adverse events based on pharmacovigilance spontaneous reporting systems

ABSTRACT Background Isavuconazole is a novel second-generation triazole antifungal agent.We conducted a retrospective investigation to analyze the occurrence of adverse events (AEs) associated with isavuconazole in real-world clinical use by data-mining self-reporting databases, aiming to provide comprehensive data for its safe and effective clinical application. Research design and methods All data obtained from the VigiBase, the EudraVigilance and the FAERS database. In the FAERS database, data mining algorithms, including reporting odds ratio (ROR) and proportional reporting ratio (PRR), were utilized to detect adverse event (AE) signals. Results A total of 4,105 AE reports related to isavuconazole were retrieved from these databases. The highest proportion of AE reports was observed in individuals aged between 18–64 years old. The general disorders and administration site conditions system organ class accounted for the largest number of reported cases. The system organ classes with most signals and reports included general disorders and administration site conditions, blood and lymphatic system disorders as well as investigations related to hepatobiliary disorders. Notably high ROR and PRR values were observed for AE signals associated with hepatobiliary disorders. Conclusion Isavuconazole has fewer interactions and less liver, kidney and cardiac toxicity than other triazole antibiotics. The identification of novel AEs necessitates meticulous attention and further investigation and analysis.

Optimisasi Penilaian Kinerja Karyawan PT. Tolan Tiga Indonesia Estate Perlabian Dengan Algoritma C4.5

In the face of fierce competition in the current global era, companies are required to prepare and always adjust their strategies to the changes that occur so that the company remains able to compete and survive. Employees who are in a company are workers which is the most important asset that must be owned and indispensable in the company and of course must be considered by all parties in order to create good performance as well as have goals to be achieved in the assessment of employee performance at PT. There Are Three Types Of Indonesian Real Estate. Employee performance appraisal in the company is seen as the driving force of the company because human resources play an active role in the running of an organization or company and the decision-making process. Machine learning tools used in predicting the assessment, using the C4.5 algorithm, the data obtained is more accurate. Machine learning is an artificial intelligence that can process data that is useful for consideration in making decisions and solving problems. C4.5 algorithm is one of the algorithms in data mining that serves to classify a class. This algorithm is a development of the ID3 algorithm. How the C4.5 algorithm works by forming a decision tree to produce a decision.

DIAFM: An Improved and Novel Approach for Incremental Frequent Itemset Mining

Traditional approaches to data mining are generally designed for small, centralized, and static datasets. However, when a dataset grows at an enormous rate, the algorithms become infeasible in terms of huge consumption of computational and I/O resources. Frequent itemset mining (FIM) is one of the key algorithms in data mining and finds applications in a variety of domains; however, traditional algorithms do face problems in efficiently processing large and dynamic datasets. This research introduces a distributed incremental approximation frequent itemset mining (DIAFM) algorithm that tackles the mentioned challenges using shard-based approximation within the MapReduce framework. DIAFM minimizes the computational overhead of a program by reducing dataset scans, bypassing exact support checks, and incorporating shard-level error thresholds for an appropriate trade-off between efficiency and accuracy. Extensive experiments have demonstrated that DIAFM reduces runtime by 40–60% compared to traditional methods with losses in accuracy within 1–5%, even for datasets over 500,000 transactions. Its incremental nature ensures that new data increments are handled efficiently without needing to reprocess the entire dataset, making it particularly suitable for real-time, large-scale applications such as transaction analysis and IoT data streams. These results demonstrate the scalability, robustness, and practical applicability of DIAFM and establish it as a competitive and efficient solution for mining frequent itemsets in distributed, dynamic environments.

Effectiveness of the Implementations of a Data Mining Algorithm Based Peak Restriction Method

The rapid growth of data in various domains has necessitated the development of advanced algorithms to efficiently analyze and manage large datasets. One of the key challenges in data mining is dealing with peak values or outliers, which can distort the analysis and lead to inaccurate predictions or decisions. This paper investigates the effectiveness of a data mining-based peak restriction method designed to minimize the influence of extreme values in large datasets. The proposed method uses a combination of data preprocessing techniques and machine learning algorithms to identify and limit peak values, enhancing the robustness and accuracy of predictive models. We evaluate the method across different datasets, including time-series, transactional, and sensor data, to assess its generalizability and performance. The results demonstrate significant improvements in model accuracy and stability, especially in applications sensitive to outliers, such as anomaly detection and predictive maintenance. Our findings suggest that the peak restriction method is a promising tool for improving the quality of data-driven models, with potential applications in various industries including finance, healthcare, and IoT. Furthermore, we discuss the limitations of the approach and propose future directions for further optimization.

Optimizing coastal groundwater quality predictions: A novel data mining framework with cross-validation, bootstrapping, and entropy analysis.

Investigating the potential of novel data mining algorithms (DMAs) for modeling groundwater quality in coastal areas is an important requirement for groundwater resource management, especially in the coastal region of Bangladesh where groundwater is highly contaminated. In this work, the applicability of DMA, including Gaussian Process Regression (GPR), Bayesian Ridge Regression (BRR) and Artificial Neural Network (ANN), for predicting groundwater quality in coastal areas was investigated. The optuna-based optimized hyperparameter is proposed to improve the accuracy of the models, including optuna-GPR and optuna-BRR as benchmark models. Combined cross-validation (CV) and bootstrapping (B) methods were used to build six predictive models. The entropy-based coastal groundwater quality index (ECWQI) was converted into a normalized index (ECWQIn), which was divided into five classes from very poor to excellent. The self-organizing map (SOM), spatial autocorrelation and fuzzy logic model were used to identify spatial groundwater quality patterns based on 12 physicochemical variables collected from 67 groundwater wells. The SOM analysis identified four distinct spatial patterns, including EC-TDS-Cl-, MgpH, Ca2+K+NO₃-, and HCO₃-SO₄2-Na+F-. The results showed that both the ANN (CV) and ANN (B) models performed better than other optuna-based models during the test phase (RMSE=0.041, MAE=0.026, R2=0.971, RAE=0.15=21 and CC=0.986) and (RMSE=0.041, MAE=0.025, R2=0.969, RAE=0.119 and CC=0.975), respectively. SO42-, Cl- and F- played an important role in the prediction accuracy. F- and SO42- showed higher spatial autocorrelation, which affected groundwater quality degradation. In addition, the ANN (CV) and ANN (B) models showed a Gaussian distribution of model errors (small standard error, <1%), indicating the stability of the model. These results indicate the efficiency of the ANN model in predicting groundwater quality in coastal areas, which would help regional water managers in real-time monitoring and management of sustainable groundwater resources.

Assessment of phenotypic diversity and morphometry of indigenous chickens in Kabwe District, Zambia

This study characterized Zambian indigenous chickens' phenotypes and morphometric traits in Kabwe District, Zambia, to create a base for breeding program development. To achieve this, 206 hens and 228 cocks we sampled in Kabwe district. Each sampled chicken was phenotyped for 4 qualitative traits: ear lobe color, shank color, comb type, and eye color. Individual linear body measurements were recorded from each chicken using a tailor's tape in centimeters, which included body length (BL), chest circumference (CC), corpus length (CL), keel length (KL), shank circumference (SC), shank length (SL), thigh circumference (TC), and thigh length (TL). The findings showed that the predominant phenotypes were single comb type (98.91%), orange eyes (80.44%), white shanks (33.70%), and red-white earlobe color (50%). The average body weight was 1698.24 g, which showed a significant positive correlation (p&lt;0.05) with all linear body measurements. For both sexes, body weight correlated positively with all linear measurements, except in males where body length showed a nonsignificant negative correlation. A stepwise regression analysis indicated that chest circumference, keel length, body length, shank circumference, thigh circumference, and corpus length (CL) are significant predictors of body weight (R² = 76.90%, P &lt; 0.05). Additionally, the CHAID data mining algorithm identified keel length as the primary predictor of body weight, with corpus length and thigh length also contributing, achieving an R² of 61.20%. These findings provide valuable insights for instituting chicken breeding programs aimed at enhancing productivity and selection decisions.

Space-time tree: a spatiotemporal construct for efficient similarity matrix calculations among network-constrained trajectories

Data mining of network-constrained trajectories has broad applications in the GIScience field. The calculation of a complete trajectory similarity matrix is a key step in various data mining algorithms. However, computing this matrix is computationally intensive for large datasets, as it involves numerous point-to-point shortest-path (PPSP) queries. To tackle this issue, we propose a new spatiotemporal construct called the space-time tree, which directly delineates the network distance from a query trajectory to any network space-time point. By constructing the space-time tree, we can efficiently compute the trajectory similarity matrix without additional PPSP queries. The space-time tree supports several similarity metrics, including closest pair distance, furthest pair distance, longest common subsequence (LCSS), and distance-weighted LCSS. It can further integrate with advanced spatiotemporal query techniques for scalable partial trajectory similarity matrix calculations. A case study using real datasets was conducted to apply the space-time tree in the trajectory clustering application. The results show that the space-time tree completed the clustering task on 0.5 million trajectories within 49 minutes, achieving a nearly 147-fold speedup compared to state-of-the-art methods.

Interaction effect between data discretization and data resampling for class-imbalanced medical datasets

Background Data discretization is an important preprocessing step in data mining for the transfer of continuous feature values to discrete ones, which allows some specific data mining algorithms to construct more effective models and facilitates the data mining process. Because many medical domain datasets are class imbalanced, data resampling methods, including oversampling, undersampling, and hybrid sampling methods, have been widely applied to rebalance the training set, facilitating effective differentiation between majority and minority classes. Objective Herein, we examine the effect of incorporating both data discretization and data resampling as steps in the analytical process on the classifier performance for class-imbalanced medical datasets. The order in which these two steps are carried out is compared in the experiments. Methods Two experimental studies were conducted, one based on 11 two-class imbalanced medical datasets and the other using 3 multiclass imbalanced medical datasets. In addition, the two discretization algorithms employed are ChiMerge and minimum description length principle (MDLP). On the other hand, the data resampling algorithms chosen for performance comparison are Tomek links undersampling, synthetic minority oversampling technique (SMOTE) oversampling, and SMOTE–Tomek hybrid sampling algorithms. Moreover, the support vector machine (SVM), C4.5 decision tree, and random forest (RF) techniques were used to examine the classification performances of the different approaches. Results The results show that on average, the combination approaches can allow the classifiers to provide higher area under the ROC curve (AUC) rates than the best baseline approach at approximately 0.8%–3.5% and 0.9%–2.5% for twoclass and multiclass imbalanced medical datasets, respectively. Particularly, the optimal results for two-class imbalanced datasets are obtained by performing the MDLP method first for data discretization and SMOTE second for oversampling, providing the highest AUC rate and requiring the least computational cost. For multiclass imbalanced datasets, performing SMOTE or SMOTE–Tomek first for data resampling and ChiMerge second for data discretization offers the best performances. Conclusions Classifiers with oversampling can provide better performances than the baseline method without oversampling. In contrast, performing data discretization does not necessarily make the classifiers outperform the baselines. On average, the combination approaches have potential to allow the classifiers to provide higher AUC rates than the best baseline approach.

Investigation of factors affecting fresh herbage yield in pea (Pisum arvense L.) using data mining algorithms.

This study was carried out to determine the factors affecting the wet grass yield of pea plants grown in Turkey. Wet grass yield was predicted using parameters such as genotype, crude protein, crude ash, acid detergent fiber (ADF), and neutral detergent fiber (NDF) with some data mining algorithms. These techniques provided easily interpretable data trees and precise cutoff values. This led to a comparison of the predictive abilities of data mining methods, including multivariate adaptive regression spline (MARS), Chi-square automatic interaction detection (CHAID), classification and regression tree (CART), and artificial neural network (ANN). To test the compatibility of the data mining algorithms, seven goodness-of-fit criteria were used. The predictive abilities of the fitted models were assessed using model fit statistics such as the coefficient of determination (R 2), adjusted R 2, root mean square error (RMSE), mean absolute percentage error (MAPE), standard deviation ratio (SD ratio), Akaike information criterion (AIC), and corrected Akaike information criterion (AICc). With the greatest R 2 and adjusted R 2 values (0.998 and 0.986) and the lowest values of RMSE, MAPE, SD ratio, AIC, and AICc (10.499, 0.7365, 0.047, 268, and 688, respectively), the MARS method was determined to be the best model for quantifying plant fresh herbage yield. In estimating the fresh herbage production of the pea plant, the results showed that the MARS method was the most appropriate model and a good substitute for other data mining techniques.

Analysis of Economic Development Mechanism Based on Stochastic Dynamic Optimization Algorithm

With the continuous improvement of China’s economic level, the research of economic data management has been paid more and more attention. As the high-tech industry plays an increasingly important role in the world economy, it has become a strategic highland for all countries in the world. How to design an economic data management model that integrates data mining techniques to achieve deep data mining and efficient utilization has become a research focus. In order to study the technical selection model of economic data, a new economic data management model based on a data mining algorithm is proposed. In the study of economic management data, the optimization model of data management is designed based on computer technology. In the optimization of data classification, a data mining algorithm is adopted to fully realize the reasonable management of data. According to the principle of the Random Dynamic Optimization Algorithm, the algorithm is improved, and its convergence speed is accelerated. The improved algorithm is applied to solve the one-dimensional cutting problem. Compared with other algorithms, the conversion method greatly improves the utilization rate of raw materials. Then, the efficiency and adaptability of the algorithm are tested by determining the algorithm and applying it to the knapsack problem.

Optimizing Procurement Efficiency by Implementing K-Means and Random Forest in Kopegtel Samarinda’s Warehouse System

Procurement is a company’s activity to purchase goods or equipment needed in operations. In the management process, a procurement management system is often used to facilitate this management, such as at CV Indocitra Multi Artha, which uses the “Sistem Warehouse Kopegtel Samarinda.” The system provides significant assistance to the company, but large requests can be overwhelming to be handled by the manager and can cause an overload information problem. Research was conducted to deal with these problems by implementing a data mining algorithm as a procurement recommendation system. K-means and Random Forest algorithms were chosen as methods for the research. The algorithm is processed within two critical steps, first by K-Means to get cluster data and then by predicting it with Random Forest to get a recommendation for whether the object should be bought or not. Hyperparameter tuning was performed to optimize the model’s performance, yielding an F1-Score of 86.95%, representing the balance between precision and recall, and an ROC AUC value of 82.34%. These substantial metric outcomes indicate that the model can provide practical recommendations

Predicting Students’ Academic Performance Using Data Mining Method

Predicting students' academic performance is a critical task in educational institutions to identify students likely to succeed or need intervention. This study leverages data mining techniques to analyze and predict student outcomes, contributing to more informed decision-making in educational management. Using data from various academic sources, the research implements three popular data mining algorithms—Naive Bayes, Multilayer Perceptron, and C4.5 Decision Tree—to classify student success. These techniques allow the identification of patterns and dependencies in student data, providing insights that can assist educators in creating tailored interventions. The results indicate that the Naive Bayes algorithm outperforms other classifiers in terms of prediction accuracy, making it a viable tool for predicting academic performance. This research underscores the potential of data mining to enhance educational outcomes by allowing proactive responses to students' academic needs.

Best IDEAS: Special Issue of the International Database Engineered Applications Symposium

Database engineered applications cover a broad range of topics including various design and maintenance methods, as well as data analytics and data mining algorithms and learning strategies for enterprise, distributed, or federated data stores [...]

Strategic Transformation of E-Commerce Big Data Classification and Mining Algorithms Based on Artificial Intelligence Era

Strategic Transformation of E-Commerce Big Data Classification and Mining Algorithms Based on Artificial Intelligence Era

An approach based on data mining and genetic algorithm to optimizing time series clustering for efficient segmentation of customer behavior

In today's highly competitive market, organizations face significant challenges in accurately understanding and segmenting customer behavior due to the inherently dynamic and evolving nature of customer interactions over time. Traditional customer segmentation methods often neglect these temporal variations, leading to ineffective business strategies and missed opportunities. This research addresses this critical gap by introducing an innovative time series-based approach for customer behavior segmentation. By modeling each customer's behavior as a time series capturing key metrics such as purchase frequency, transaction amounts, and customer lifecycle costs the proposed method dynamically adapts to behavioral changes over time. To enhance segmentation precision, a genetic algorithm is employed to optimize feature weights, ensuring that the most relevant factors are emphasized. These optimized features are then clustered using spectral clustering to identify distinct and meaningful customer segments. The effectiveness of the proposed method is validated using 30 months of transactional data from a payment services company. The results demonstrate that the proposed approach, particularly when combined with spectral clustering and optimally weighted features, significantly surpassing the performance of traditional static segmentation techniques. This research not only provides a more accurate framework for uncovering hidden patterns in customer behavior but also delivers actionable insights for targeted marketing and personalized customer strategies.