Linear Decision Research Articles

Intrusion Detection: A Comparison Study of Machine Learning Models Using Unbalanced Dataset

AbstractThe worldwide process of converting most activities of both corporate and non-corporate entities into digital formats is now firmly established. Machine learning models are necessary to serve as a tool for preventing illegal intrusion onto different networks. The machine learning (ML) model's strengths and drawbacks pertain to intrusion detection (IDS) tasks. This study used an experimental methodology to assess the efficacy of various ML models, including linear SVC, LR, random forest (RF), decision tree (DT), and XGBoost, in detecting intrusion on the UNSW NB15 datasets. The objective is to compare the strengths and shortcomings of these models. Data exploration, Feature engineering, selection and a test set of 15%, a validation set of 15%, and a training set of 70% respectively were used for data splitting. Performance evaluation was carried out using accuracy, recall, precision F1-score and confusion matrix plotted. The outcome of the experiment shows a percentage of 92.71% (1, normal) and 7.29% (0, attack) for normal traffic and attack traffic respectively. Performance evaluation results showed that RF and XGBoost outperformed the other ML models. Hence, ML models can effectively be used to detect system attacks. We intend to expand this research in the future and use the paradigm in a real-world setting with further conclusions and justifications.

Predicting the Multiphotonic Absorption in Graphene by Machine Learning

This study analyzes the nonlinear optical properties exhibited by graphene, focusing on the nonlinear absorption coefficient and the nonlinear refractive index. The evaluation was conducted using the Z-scan technique with a 532 nm wavelength laser at various intensities. The nonlinear optical absorption and the nonlinear optical refractive index were measured. Four machine learning models, including linear regression, decision trees, random forests, and gradient boosting regression, were trained to analyze how the nonlinear optical absorption coefficient varies with variables such as spot radius, maximum energy, and normalized minimum transmission. The models were trained with synthetic data and subsequently validated with experimental data. Decision tree-based models, such as random forests and gradient boosting regression, demonstrated superior performance compared to linear regression, especially in terms of mean squared error. This work provides a detailed assessment of the nonlinear optical properties of graphene and highlights the effectiveness of machine learning methods in this context.

Smart farming based on IoT to predict conditions using machine learning

<span>Smart farming is a type of technology that utilizes the internet of things (IoT) to provide information on agricultural and environmental conditions as well as perform automation. Some of these ecological conditions can be used and analyzed in machine learning (ML) data management. This study focuses on utilizing ML algorithms to find the best prediction; typically used methods include linear regression, decision tree (DT), random forest (RF), and extreme gradient boosting (XGBoost). In the application of smart farming, research on IoT and artificial intelligence (AI) is still uncommon since most IoT cannot make predictions like AI. Because basically, some IoT can't make predictions as AI does. In this Study, predictions were made by looking at the regression results in the form of root mean square error (RMSE) and absolute error. The results show a strong and weak correlation between features (positive or negative). The best prediction results are obtained by XGBoost when predicting temperature (RMSE 6.656 and absolute error 3.948) and (soil moisture 17.151 and absolute error 11.269). However, using different parameters (RMSE RF and absolute error DT) on RF and DT resulted in good and distinct results. Linear regression, on the other hand, produced unsatisfactory and poor result.</span>

Comparison of SEMG-Based Hand Gesture Classifiers

Machine learning techniques have shown success in classifying hand gestures. As the prevalence of prosthetic devices continues to rise, the adoption of non-invasive technologies, such as surface electromyography (sEMG), becomes paramount. This study systematically assesses the isolated influence of classification algorithms within hand gesture recognition (HGR) systems using sEMG data and dynamic time warping (DTW) based features. This approach effectively handles temporal variations in sEMG signals by leveraging DTW, ensuring input features are invariant to gesture speed. Six supervised learning classifiers were evaluated: the multilayer perceptron, support vector machine, logistic regression, linear discriminant analysis, k-nearest neighbors, and decision tree. Cross-validation was employed to fine-tune the segmentation hyperparameters, significantly improving results. To ensure reproducibility, the source code has been made available, the proposed system design has been detailed, and the evaluation protocols have been described. Our findings indicate that logistic regression outperformed other classifiers in this setup, achieving 95.2% accuracy in classifying six hand movements from ten healthy individuals, representing a 1.6% improvement over the best previously reported performance using the same publicly available dataset. Future research will assess the proposed HGR system’s generalization capability on larger datasets suitable for training more complex classifiers, including deep learning models.

PREDICTIVE TECHNIQUES AND ARTIFICIAL INTELLIGENCE MODELS IN THE MANAGEMENT OF DEFAULTY PUBLIC DEBTS: COMPARISON BETWEEN LINEAR REGRESSION AND DECISION TREES

The study analyzed focuses on the application of predictive models, specifically Linear Regression and Decision Trees, for the management of defaulted debts in the public context of the United States. The main objective of the work is to compare the effectiveness of these models in predicting the compliance of debts with more than 120 days, assisting in directing these debts to the Treasury Offset Program (TOP), an essential initiative for the government's financial recovery. The problem that the study addresses is the need for effective management of defaulted public debts, seeking to ensure compliance with public financial policies that promote compliance and the adequate redirection of financial resources to the government. This is particularly important to ensure fiscal transparency and accountability of federal agencies. The methodology used in the study was quantitative, based on the analysis of data on eligible debts extracted from reports of the US Treasury. Linear Regression and Decision Tree models were applied, with performance metrics such as Mean Absolute Error (MAE), Mean Squared Error (MSE), Root Mean Squared Error (RMSE) and Coefficient of Determination (R²). The study addressed financial and temporal variables to analyze the behavior of these debts and their compliance. The main results show that both models presented high accuracy in predictions, with Linear Regression showing a perfect fit (R² = 1) and Decision Trees standing out in capturing non-linear nuances of the data. The variable "Compliance Rate Amount" was identified as the most significant in the Decision Tree model, suggesting that the amount of the compliance rate is one of the most important factors in predicting the compliance of defaulted debts. This study offers valuable contributions to the field of public management, by demonstrating that the use of predictive models can help optimize debt recovery, improve fiscal transparency and contribute to more informed decision-making.

The Comprehensive Investigation for Covid-19 Trend Prediction Through Machine Learning and Deep Learning

Corona Virus Disease (Covid-19) has surely been a challenging problem to solve for the past few years. Due to the diversity in the form of dataset, it is essential to obtain accurate predictive results of Covid-19 trends. This paper analyzes different artificial intelligence methods used in Covid-19 trend prediction, including several machine learning and deep learning methods. More specifically, this work investigates linear regression, random forest, and decision trees in terms of machine learning and delves into Artificial Neural Network (ANN) as well as Long Short-Term Memory (LSTM) for deep learning. By comparing various past works, the effectiveness of machine learning and deep learning methods is achieved by their hidden algorithms, such as the Multiple Linear Regression (MLR) model for linear regression analysis. Incorporation with other models or methods is applied in deep learning. For example, Ensemble Empirical Mode Decomposition (EEMD) is included in ANN structure to decrease the noises within the Covid-19 datasets. Furthermore, the paper also inquiries into potential improvement of some drawbacks in predictive results for Covid-19 trends by reviewing related works of expert system and transfer learning as well as domain adaptation. The machine learning and deep learning models could provide accurate predictive results as a reference for related organizations to consider or establish insightful policies.

A Multi-Model Output Fusion Strategy Based on Various Machine Learning Techniques for Product Price Prediction

In the digital era, precise product price prediction becomes crucial for enhancing competitiveness in the online marketplace. This paper presents a hybrid model framework that enhances the accuracy of online product price predictions by integrating several machine learning algorithms, including Linear Regression, Decision Trees, and Gradient Boosting. The objective of this approach is to leverage the distinct advantages of each model to address their individual limitations and create a robust unified predictive model. This integration allows for improved handling of complex data relationships and diverse market dynamics that are typical in online sales environments. The results demonstrate that the hybrid model achieves superior prediction accuracy, as reflected in reduced Mean Absolute Error (MAE) and Root Mean Square Error (RMSE) metrics, and an exceptionally high R ² value compared to single-model approaches. These outcomes underscore the efficacy of combining multiple predictive models to enhance the precision of price forecasts in the highly competitive online marketplace. This model fusion strategy not only provides more accurate pricing predictions but also offers strategic insights into the optimal pricing strategies for businesses looking to enhance their market position.

Comparative Study of Fast Optimization Method for Four‐Intensity Measurement‐Device‐Independent Quantum Key Distribution Through Machine Learning

AbstractThe four‐intensity protocol for measurement‐device‐independent (MDI) quantum key distribution (QKD) is renowned for its excellent performance and extensive experimental implementation. To enhance this protocol, a machine learning‐driven rapid parameter optimization method is developed. This initial step involved a speed‐up technique that quickly pinpoints the worst‐case scenarios with minimal data points during the optimization phase. This is followed by a detailed scan in the key rate calculation phase, streamlining data collection to fit machine learning timelines effectively. Several machine learning models are assessed—Generalized Linear Models (GLM), k‐Nearest Neighbors (KNN), Decision Trees (DT), Random Forests (RF), XGBoost (XGB), and Multilayer Perceptron (MLP)—with a focus on predictive accuracy, efficiency, and robustness. RF and MLP were particularly noteworthy for their superior accuracy and robustness, respectively. This optimized approach significantly speeds up computation, enabling complex calculations to be performed in microseconds on standard personal computers, while still achieving high key rates with limited data. Such advancements are crucial for deploying QKD under dynamic conditions, such as in fluctuating fiber‐optic networks and satellite communications.

Interpretable Machine Learning: A Case Study on Predicting Fuel Consumption in VLGC Ship Propulsion

The integration of machine learning (ML) in marine engineering has been increasingly subjected to stringent regulatory scrutiny. While environmental regulations aim to reduce harmful emissions and energy consumption, there is also a growing demand for the interpretability of ML models to ensure their reliability and adherence to safety standards. This research highlights the need to develop models that are both transparent and comprehensible to domain experts and regulatory bodies. This paper underscores the importance of transparency in machine learning through a use case involving a VLGC ship two-stroke propulsion engine. By adhering to the CRISP-DM standard, we fostered close collaboration between marine engineers and machine learning experts to circumvent the common pitfalls of automated ML. The methodology included comprehensive data exploration, cleaning, and verification, followed by feature selection and training of linear regression and decision tree models that are not only transparent but also highly interpretable. The linear model achieved an RMSE of 23.16 and an MRAE of 14.7%, while the accuracy of decision trees ranged between 96.4% and 97.69%. This study demonstrates that machine learning models for predicting propulsion engine fuel consumption can be interpretable, adhering to regulatory requirements, while still achieving adequate predictive performance.

Estimation Stand Volume, Basal Area and Quadratic Mean Diameter Using Landsat 8 OLI and Sentinel‐2 Satellite Image With Different Machine Learning Techniques

ABSTRACTThe data required for sustainable forest planning is provided by traditional forest inventories, which are labor, time, and cost‐intensive. Providing this data quickly, reliably, and accurately is crucial for planners and researchers. The objective of this study was to predict stand basal area (BA), stand volume (V), and quadratic mean diameter (dq) by leveraging vegetation indices (VIs) and reflectance (R) derived from Landsat 8 OLI and Sentinel 2 satellite images, along with topographic (T) data obtained from ALOS‐PALSAR satellite imagery. Forest inventory data for a total of 250 sample plots were used for modeling in the study. Stand parameters were estimated using support vector machines (SVM), multiple linear regression (MLR), decision tree (DT), and random forest (RF) algorithms. In modeling V, BA, and dq, both individual and combinations of R, VIs, and T values obtained from satellite imagery were used as independent variables. Using the generated datasets, each of the stand parameters was modeled separately with MLR, SVM, RF, and DT algorithms, and the success of the models was compared to determine the modeling technique and dataset with the highest success for the relevant parameter. The results showed that for each stand parameter, the highest model success was achieved in the combined dataset, which was created by combining all datasets. However, in terms of modeling techniques, the highest success for each stand parameter was achieved with different modeling techniques. The highest success for V is obtained in the model using the SVM method (R2 = 0.78; RMSE = 0.28 m3/ha), the RF method yielded the highest model performance for BA (R2 = 0.70; RMSE = 2.53 m2/ha), and finally, the highest success for dq was obtained in the DT method (R2 = 0.74; RMSE = 0.02 cm). In general, the datasets obtained from Sentinel 2 images showed higher model success than the datasets obtained from Landsat 8 OLI images.

MRI-Based Machine Learning for Prediction of Clinical Outcomes in Primary Central Nervous System Lymphoma.

A portion of individuals diagnosed with primary central nervous system lymphomas (PCNSL) may experience early relapse or refractory (R/R) disease following treatment. This research explored the potential of MRI-based radiomics in forecasting R/R cases in PCNSL. Forty-six patients with pathologically confirmed PCNSL diagnosed between January 2008 and December 2020 were included in this study. Only patients who underwent pretreatment brain MRIs and complete postoperative follow-up MRIs were included. Pretreatment contrast-enhanced T1WI, T2WI, and T2 FLAIR imaging were analyzed. A total of 107 radiomic features, including 14 shape-based, 18 first-order statistical, and 75 texture features, were extracted from each sequence. Predictive models were then built using five different machine learning algorithms to predict R/R in PCNSL. Of the included 46 PCNSL patients, 20 (20/46, 43.5%) patients were found to have R/R. In the R/R group, the median scores in predictive models such as support vector machine, k-nearest neighbors, linear discriminant analysis, naïve Bayes, and decision trees were significantly higher, while the apparent diffusion coefficient values were notably lower compared to those without R/R (p < 0.05). The support vector machine model exhibited the highest performance, achieving an overall prediction accuracy of 83%, a precision rate of 80%, and an AUC of 0.78. Additionally, when analyzing tumor progression, patients with elevated support vector machine and naïve Bayes scores demonstrated a significantly reduced progression-free survival (p < 0.05). These findings suggest that preoperative MRI-based radiomics may provide critical insights for treatment strategies in PCNSL.

Data‐driven prediction of prolonged air leak after video‐assisted thoracoscopic surgery for lung cancer: Development and validation of machine‐learning‐based models using real‐world data through the ePath system

AbstractIntroductionThe reliability of data‐driven predictions in real‐world scenarios remains uncertain. This study aimed to develop and validate a machine‐learning‐based model for predicting clinical outcomes using real‐world data from an electronic clinical pathway (ePath) system.MethodsAll available data were collected from patients with lung cancer who underwent video‐assisted thoracoscopic surgery at two independent hospitals utilizing the ePath system. The primary clinical outcome of interest was prolonged air leak (PAL), defined as drainage removal more than 2 days post‐surgery. Data‐driven prediction models were developed in a cohort of 314 patients from a university hospital applying sparse linear regression models (least absolute shrinkage and selection operator, ridge, and elastic net) and decision tree ensemble models (random forest and extreme gradient boosting). Model performance was then validated in a cohort of 154 patients from a tertiary hospital using the area under the receiver operating characteristic curve (AUROC) and calibration plots.ResultsTo mitigate bias, variables with missing data related to PAL or those with high rates of missing data were excluded from the dataset. Fivefold cross‐validation indicated improved AUROCs when utilizing key variables, even post‐imputation of missing data. Dichotomizing continuous variables enhanced performance, particularly when fewer variables were employed in the decision tree ensemble models. Consequently, regression models incorporating seven key variables in complete case analysis demonstrated superior discriminatory ability for both internal (AUROCs: 0.77–0.84) and external cohorts (AUROCs: 0.75–0.84). These models exhibited satisfactory calibration in both cohorts.ConclusionsThe data‐driven prediction model implementing the ePath system exhibited adequate performance in predicting PAL post‐video‐assisted thoracoscopic surgery, optimizing variables and considering population characteristics in a real‐world setting.

Combinatorial Green Optimization of Individual Quick Freezer for Energy Savings through Linear Programming and Decision Theory of Shrimp Processing Industry : A Mathematical Approach

Combinatorial Green Optimization of Individual Quick Freezer for Energy Savings through Linear Programming and Decision Theory of Shrimp Processing Industry : A Mathematical Approach

Prediction of room temperature in Trombe solar wall systems using machine learning algorithms

A Trombe wall-heating system is used to absorb solar energy to heat buildings. Different parameters affect the system performance for optimal heating. This study evaluated the performance of four machine learning algorithms—linear regression, k-nearest neighbors, random forest, and decision tree—for predicting the room temperature in a Trombe wall system. The accuracy of the algorithms was assessed using R² and RMSE values. The results demonstrated that the k-nearest neighbors and random forest algorithms exhibited superior performance, with R² and RMSE values of 1 and 0. In contrast, linear regression and decision tree showed weaker performance. These findings highlight the potential of advanced machine learning algorithms for accurate room temperature prediction in Trombe wall systems, enabling informed design decisions to enhance energy efficiency.

Software data objects application integrity modeling in medication dispensing errors predictions using machine learning algorithms in workflow-based pharmacy software systems

The study focuses on using the software objects created at each stage of the workflow-based systems (and in our article we can focus on the pharmacy-based medication dispensing software systems) to predict the anomalies or potential issues in that software object constructed at each stage of the workflow using the historical issues recorded in that specific workflow stage using linear regression, decision trees, and neural network algorithms. By utilizing historical data and integrity metrics, the research aims to forecast potential failures and maintain the reliability of software applications. The focus is on employing algorithms such as linear regression, decision trees, and neural networks to enhance predictive accuracy and facilitate informed decisions or recommendations to the workflow users (pharmacists) about the potential medication dispensing error the software objects constructed at that state or workflow could cause and advise to take necessary corrections.

An efficient approach for EMG controlled pattern recognition system based on MUAP identification and segregation

An Electromyography (EMG) based pattern recognition system constitutes various steps of signal processing and control engineering from signal acquisition to real-time control. Efficient control of external devices largely depends on the signal processing steps executed before the final output. This work presents a new approach to signal processing using Motor Unit Action Potential (MUAP) based signal decomposition and segmentation. An MUAP is a neurological response during muscle contraction. Due to the higher contact area of surface electrodes, MUAPs from multiple muscles are captured. An MUAP generated from a single muscle usually has identical waveshapes and similar discharging rates and usually lasts for 8–15 ms. These are known as primary MUAPs. The proposed algorithm identifies and uses the primary observed MUAPs for feature extraction and classification. Firstly, noise signals are eliminated by a determined noise margin, which also separates the active muscle movement signals. Next, a novel MUAP identification algorithm is implemented to detect the MUAP trains. Then, identified primary MUAPs are used to make segments with variable widths to extract feature vectors. Based on the correlation score of all the primary MUAPs, the segmentation is performed, which results in segmentation width varying from 110–200 ms. The achieved segmentation width is lesser than the conventional overlapping and non-overlapping methods — the proposed approach results in a 20 to 50% reduction in the segmentation width. Four different classifiers are tested during the machine learning stage to investigate the performance of the proposed approach. The obtained feature sets are then used to train the Linear Discriminant Analysis (LDA), K-Nearest Neighbor (kNN), Decision Tree (DT), and Random Forest (RF) classifiers. The classifiers are tested with precision, recall, F1 score, and accuracy. The kNN and DT classifiers performed better than the LDA and RF classifiers. The maximum precision and recall are 100% while the maximum achieved accuracy is 98.56%. The comparative results show higher accuracy even at lower segmentation widths than the conventional constant window scheme. The kNN and DT classifiers provide a 5% to 15% increment in accuracy compared to the constant window segmentation-based approach.

Development of a unified framework of low-rank approximation and deep neural networks for predicting the spatial variability of SSC in `Spania' watermelons using vis/NIR hyperspectral imaging

Soluble Solids Content (SSC) is an important quality attribute that represents the internal quality of fruits. Visible and near-infrared (Vis/NIR) combined with chemometric algorithms are now popular methods for non-invasive measurement and visualization of SSC. However, in fruits with a thick rind and a large flesh volume, such as watermelon, SSC is not evenly distributed across the fruit. The variability in SSC across the fruit flesh may have an adverse effect on the accuracy of quality analysis algorithms. Thus, this paper presents an accurate and efficient approach for predicting the spatial variation of SSC in watermelons using a combined framework of low-rank approximation and deep neural networks. Watermelon ‘Spania’ was selected for this study, and hyperspectral images of each watermelon were taken from various views, including the top, bottom, and two lateral views. The low-rank property is employed as a constraint in this approach to eliminate the unwanted variations in the spectral data. The low-rank component of the spectral data, free of unwanted variations, is then fed into a fully connected neural network (FNN) for the prediction of watermelon SSC values. The proposed approach obtained optimal performance in calibration and prediction with RC2=0.982, RMSEP = 0.132, and RP2=0.945, RMSEP = 0.195 respectively. Further, the prediction outcomes of the proposed method were compared with state-of-the-art such as Partial Least Square Regression (PLSR), Support Vector Regression (SVR), Multiple Linear Regression (MLR), Decision Tree (DT), and Random Forest (RF). The overall results showed that combining HSI data with a low-rank and deep neural network framework is an efficient method for accurately predicting SSC in watermelons as well as correctly predicting spatial variability of SSC in individual fruits.

Parametric Analysis of Critical Buckling in Composite Laminate Structures under Mechanical and Thermal Loads: A Finite Element and Machine Learning Approach.

This research focuses on investigating the buckling strength of thin-walled composite structures featuring various shapes of holes, laminates, and composite materials. A parametric study is conducted to optimize and identify the most suitable combination of material and structural parameters, ensuring the resilience of structure under both mechanical and thermal loads. Initially, a numerical approach employing the finite element method is used to design the C-section thin-walled composite structure. Later, various structural and material parameters like spacing ratio, opening ratio, hole shape, fiber orientation, and laminate sequence are systematically varied. Subsequently, simulation data from numerous cases are utilized to identify the best parameter combination using machine learning algorithms. Various ML techniques such as linear regression, lasso regression, decision tree, random forest, and gradient boosting are employed to assess their accuracy in comparison with finite element results. As a result, the simulation model showcases the variation in critical buckling load when altering the structural and material properties. Additionally, the machine learning models successfully predict the optimal critical buckling load under mechanical and thermal loading conditions. In summary, this paper delves into the study of the stability of C-section thin-walled composite structures with holes under mechanical and thermal loading conditions using finite element analysis and machine learning studies.

Noninvasive brain stimulation during EEG improves machine learning classification in chronic stroke.

In individuals with chronic stroke and hemiparesis, noninvasive brain stimulation (NIBS) may be used as an adjunct to therapy for improving motor recovery. Specific states of movement during motor recovery are more responsive to brain stimulation than others, thus a system that could auto-detect movement state would be useful in correctly identifying the most effective stimulation periods. The aim of this study was to compare the performance of different machine learning models in classifying movement periods during EEG recordings of hemiparetic individuals receiving noninvasive brain stimulation. We hypothesized that transcranial direct current stimulation, a form of NIBS, would modulate brain recordings correlating with movement state and improve classification accuracies above those receiving sham stimulation. Electroencephalogram data were obtained from 10 participants with chronic stroke and 11 healthy individuals performing a motor task while undergoing transcranial direct current stimulation. Eight traditional machine learning algorithms and five ensemble methods were used to classify two movement states (a hold posture and an arm reaching movement) before, during and after stimulation. To minimize compute times, preprocessing and feature extraction were limited to z-score normalization and power binning into five frequency bands (delta through gamma). Classification of disease state produced significantly higher accuracies in the stimulation (versus sham) group at 78.9% (versus 55.6%, p < 0.000002). We observed significantly higher accuracies when classifying stimulation state in the chronic stroke group (77.6%) relative to healthy controls (64.1%, p < 0.0095). In the chronic stroke cohort, classification of hold versus reach was highest during the stimulation period (75.2%) as opposed to the pre- and post-stimulation periods. Linear discriminant analysis, logistic regression, and decision tree algorithms classified movement state most accurately in participants with chronic stroke during the stimulation period (76.1%). For the ensemble methods, the highest classification accuracy for hold versus reach was achieved using low gamma frequency (30-50 Hz) as a feature (74.5%), although this result did not achieve statistical significance. Machine learning algorithms demonstrated sufficiently high movement state classification accuracy in participants with chronic stroke performing functional tasks during noninvasive brain stimulation. tDCS improved disease state and movement state classification in participants with chronic stroke.

OPTIMIZING RETAIL DEMAND FORECASTING: A PERFORMANCE EVALUATION OF MACHINE LEARNING MODELS INCLUDING LSTM AND GRADIENT BOOSTING

Effective demand forecasting is vital for inventory management in retail. This study evaluates five machine learning models—Linear Regression (LR), Decision Tree Regressor (DTR), Random Forest Regressor (RFR), Gradient Boosting (GB), and Long Short-Term Memory (LSTM)—for predicting retail demand. Utilizing a dataset with transactional sales, promotions, calendar events, and external factors like weather and economic indicators, we applied rigorous preprocessing and feature engineering. Performance was assessed using Mean Absolute Error (MAE), Root Mean Square Error (RMSE), and R-squared (R²). Results show that LSTM outperforms other models with an MAE of 9.53, RMSE of 14.67, and R² of 0.90, excelling in capturing temporal dependencies and complex demand patterns. Gradient Boosting and Random Forest also performed well, while Linear Regression and Decision Tree Regressor showed limitations. This study highlights the effectiveness of advanced models, particularly LSTM, for enhancing demand forecasting accuracy and offers valuable insights for optimizing retail inventory and operations.