جستجوی مقالات مرتبط با کلیدواژه "machine learning" در نشریات گروه "صنایع"

This study addresses the Hub Location-Routing Problem (HLRP) in transportation networks, considering the inherent uncertainty in travel times between nodes. We employed a method centered on data-driven robust optimization, utilizing Support Vector Clustering (SVC) to form an uncertainty set grounded in empirical data. The proposed methodology is compared against traditional uncertainty sets, showcasing its superior performance in providing robust solutions. A comprehensive case study on a retail store's transportation network in Tehran is presented, demonstrating significant differences in hub locations, allocations, and vehicle routes between deterministic and robust models. The SVC-based model proves to be particularly effective, yielding substantially improved objective function values compared to polyhedral and box uncertainty sets. The study concludes by highlighting the practical significance of this research and suggesting future directions for advancing transportation network optimization under uncertainty.

One of the constant problems that people with mental health conditions are faced with now is that they cannot establish a good relationship with their therapist, or the client's disease type is not in the therapist's specialty. These clients may not receive adequate treatment and stop the therapy before feeling well. Therefore, the classification of mental patients based on their disorder types and allocating a therapist with the same expertise to them could lead to better treatment and improve the quality of the therapy sessions. This paper will compare several machine learning (ML) algorithms to classify patients with mental conditions. Moreover, benefiting from the best ML algorithm, patients will be categorized into different classes based on their disorder types. Finally, a mathematical model will be developed to determine the allocation policy of therapists to each group of patients to maximize the summation of the utilization between therapists and patients. To explore the implementation of the proposed method, we have conducted a real-life case study to assess the validation of the model.

Customer churn is one of the challenges of business management in today's complex competitive environment. For this purpose, the organization must have an efficient system to detect and analyze the factors influencing customer churn. To conduct this research, an attempt has been made to build a hybrid model based on data mining approaches from information related to 5830 customers of a chain store (demographic information and information based on customer purchase records) with 17 qualitative and quantitative characteristics. The features of higher importance were identified to build the model in the first stage using the logistic regression algorithm. In the second stage, the support vector machine algorithm, a critical supervised learning algorithm, was used to classify the customers and rank the essential features. Finally, the proposed model has been implemented as a case study in the chain store industry. The results indicate the optimal efficiency of the proposed analysis method. This research has been done to identify the influential factors in customer churn and focus on providing new solutions to reduce churn in the retail industry. Also, the results show that age, marital status, and average monthly income from the set of demographic features and how to get to know the store, the share of online shopping, and special sales from the set of features related to customer transaction records are among the most important factors affecting customer churn. In addition, practical suggestions have been presented that can be used for tactical and strategic planning of chain stores to attract and retain customers.

Optimizing production in the plastic extrusion industry is a pivotal task for small scale industries. To enhance the efficiency in today’s competitive market being a small-scale manufacturer over their peers is challenging. With the limited resources, having constraints on manpower, capital, space, often facing fluctuations in demand and production, simultaneously maintaining high quality became very important for the success. Among the plethora of KPIS used in manufacturing, Overall Equipment Effectiveness (OEE) stands out as corner stone. In this study, we collected real-world data from a plastic extrusion company. i.e., an HDPE Pipe manufacturing company. It serves as the backdrop for our study, this is based on the plastic extrusion sector and set out a goal of enhancing OEE through a comparative investigation of various ML models.  To forecast and estimate OEE values, we used various Machine Learning models and examine each algorithm’s performance using metrics like Mean Squared Error (MSE) and model comparisons using the Akaike Information Criterion (AIC) and Bayesian Information Criterion (BIC), creating a comprehensive picture of each algorithm’s strength which enables the small businesses to make informed decisions and empowers them to stay agile and adapt to the changes in the manufacturing environment.

We consider a continuous model of the optimal control of the customer dynamics based on marketing policies as a non-autonomous system of ODEs. The model tracks the history of the simultaneous changes from the beginning to the current time for the evolution of the company's regular, referral, and potential customers. We then present a new supervised machine-learning algorithm for the numerical simulation of the problem. The proposed learning algorithm implements a polynomial kernel to simplify the formulation of the method. To avoid computational complexity, the Bernstein kernels are used to get a simple optimization marketing strategy by using the Support Vector Regression (SVR) in a least-squares framework. Some numerical experiments are carried out to support the proposed model and the method. The method provides approximate numerical results with high accuracy by kernels of polynomials of low degree. The running time of the technique is also illustrated versus the increasing number of training points to see the polynomial behavior of the running time.

This paper considers an accurate and efficient diabetes detection scheme via machine learning. It uses the science of data mining and pattern matching in its diabetes diagnosis process. It implements and evaluates 4 machine learning classification algorithms, namely Decision tree, Random Forest, XGBoost and LGBM. Then selects and introduces the one that performs the best towards its objective using multi-criteria decision-making methods. Its results reveal that Random Forest algorithm outperformed other algorithms with higher accuracy. It also examines the details of features that have a greater effect on diabetes detection. Considering that diabetes is one of the most deadly, disabling, and costly diseases observed today, its alarmingly increasing rates, and difficulty of its diagnosis because of many vague signs and symptoms, utilization of such approach can help doctors increase accuracy of their diagnosis and treatment schemes. Hence, this paper uses the science of data mining as a tool to gather and analyze existing data on diabetes and help doctors with its diagnosis and treatment process. The main contribution of this paper can therefore be its applied nature to an essential field and accuracy of its pattern recognition via several analytical approaches.

This research aims to identify the initial OEE (Overall Equipment Efficiency) values on automated packaging machines with a 2d barcode track and trace system. Quantitative research methods used to obtain the OEE value, analysis of factors affecting the OEE values, developing a strategy to make improvements, and evaluate these strategies on the level of machine productivity. The results showed that the OEE value of the automatic packing machine with 2d barcode track and trace system was 30.49%. The value is still far below the company’s standard of 85%. Availability (av) and the rate of quality (rq) are quite high. However, the performance rate (pe) is still low until the OEE value is low. Based on six big loss analysis, the biggest loss that leads to low OEE value is decreasing speed. Machine is operated at speeds that do not match design (ideal); usually, the real speed is lower than the ideal speed, that can be caused by small product batch sizes. Based on average OEE after a problem solving of 67.67 %. An increase in value gained by 37.18 % of previous achievements. To sum up, this study contributes to the development of best practices and strategies for optimizing machine productivity and improving overall equipment effectiveness in various manufacturing contexts.

During the crisis, relief supply chain management (also known as humanitarian supply chain management) has received great attention these days. The core questions facing many humanitarian organizations are: where are their strengths/weaknesses? Are they positioned to be effective in their supply chain system? What challenges do you need to overcome? What do they need to do to take advantage of the technological opportunities offered nowadays? These questions have been addressed them extensively during the past two decades. This paper tries to review and classify some of the papers carried out in key areas of the humanitarian supply chain such as location, certainty and uncertainty, relief teams and injured (patient) classification, machine learning, queue theory, the employed research methods, solution methods, and the type of objective functions. The paper begins first to define what the “humanitarian” ecosystem may include, and which actors play important roles. After, certain critical views of the humanitarian relief supply chain are examined. The critical views of the humanitarian relief supply chain would help researchers to introduce further research orientations and areas to overcome crises in the real world.

Tuberculosis (TB) remains a significant public health concern in Europe, necessitating effective disease management and resource allocation. Predicting short-term TB incidence rates using machine learning algorithms offers a data-driven approach to aid policymakers and healthcare professionals in making informed decisions. Machine learning (ML) algorithms are essential for prediction tasks due to their ability to establish a relationship for data sequences. In this study, three machine learning algorithms, namely, Decision Tree (DT), Random Forest (RF), and Artificial Neural Network (ANN), are implemented to predict the tuberculosis incidence rates and to compare the efficacy of ML algorithms for tuberculosis incidence rates prediction for 2025, among Europe. Even though all models achieved considerable results, DT obtained superior prediction rates for the future TB incidence rate with MSE, MAE, and R2 of 0.000555, 0.01506, and 0.96430 while RF 0.000882, 0.01781, and 0.94329, and ANN 0.000767, 0.02315, and 0.95066. The prediction results showed that a significant decrease in TB incidence rates is expected for 2025 form 49,752 in 2019 to 38,509 in 2025, except Finland and Malta.

This study addresses the Hub Location-Routing Problem (HLRP) in transportation networks, considering the inherent uncertainty in travel times between nodes. We employed a method centered on data-driven robust optimization, utilizing Support Vector Clustering (SVC) to form an uncertainty set grounded in empirical data. The proposed methodology is compared against traditional uncertainty sets, showcasing its superior performance in providing robust solutions. A comprehensive case study on a retail store's transportation network in Tehran is presented, demonstrating significant differences in hub locations, allocations, and vehicle routes between deterministic and robust models. The SVC-based model proves to be particularly effective, yielding substantially improved objective function values compared to polyhedral and box uncertainty sets. The study concludes by highlighting the practical significance of this research and suggesting future directions for advancing transportation network optimization under uncertainty.

Feature selection is the process of picking the most effective feature among a considerable number of features in the dataset. However, choosing the best subset that gives a higher performance in classification is challenging. This study constructed and validated multiple metaheuristic algorithms to optimize Machine Learning (ML) models in diagnosing Alzheimer’s. This study aims to classify Cognitively Normal (CN), Mild Cognitive Impairment (MCI), and Alzheimer’s by selecting the best features. The features include Freesurfer features extracted from Magnetic Resonance Imaging (MRI) images and clinical data. We have used well-known ML algorithms for classifying, and after that, we used multiple metaheuristic methods for feature selection and optimizing the objective function of the classification. We considered the objective function a macro-average F1 score because of the imbalanced data. Our procedure not only reduces the irreverent features but also increases the classification performance. Results showed that metaheuristic algorithms could improve the performance of ML methods in diagnosing Alzheimer’s by 20%. We found that classification performance can be significantly enhanced by using appropriate metaheuristic algorithms. Metaheuristic algorithms can help find the best features for medical classification problems, especially Alzheimer’s.

Customer churn prediction has been gaining significant attention due to the increasing competition among mobile service providers. Machine learning algorithms are commonly used to predict churn; however, their performance can still be improved due to the complexity of customer data structure. Additionally, the lack of interpretability in their results leads to a lack of trust among managers. In this study, a step-by-step framework consisting of three layers is proposed to predict customer churn with high interpretability. The first layer utilizes data preprocessing techniques, the second layer proposes a novel classification model based on supervised and unsupervised algorithms, and the third layer uses evaluation criteria to improve interpretability. The proposed model outperforms existing models in both predictive and descriptive scores. The novelties of this paper lie in proposing a hybrid machine learning model for customer churn prediction and evaluating its interpretability using extracted indicators. Results demonstrate the superiority of clustered dataset versions of models over non-clustered versions, with KNN achieving a recall score of almost 99% for the first layer and the cluster decision tree achieving a 96% recall score for the second layer. Additionally, parameter sensitivity and stability are found to be effective interpretability evaluation metrics.

Selection of the most influential factors to improve the performance of organizations has consistently been a significant task for project managers. These underlying factors aim to prevent the failure of the project and to improve the performance of employees. The success of the organization's projects is directly correlated to customer satisfaction, time, cost, and product quality at the time of project completion. In this paper, after reviewing the literature on the elements influencing the project's success, the extent to which each factor affects the project's success is accessed. A practical data evaluation method to predict the most underlying item is a machine learning algorithm, a perfect contributory method for project managers to examine the influential factors. After identifying the component with the highest effect on the project success, validating the selected items in a real-world practice paves the way for assessing that factor's effectiveness. In this study, after selecting the Agile approach as the most notable, the simulation models were utilized to measure the proportion of organizational performance improvement. Agile Management, which is considered in the actual case, signifies implementing the Scrum method and all the definitions and phases related to this method in the organization. The analyzed Agile practice (Scrum) for the case study decremented the project cost and time substantially and enhanced the service and product quality.

The Markov chain is widely used in state-dependent inventory control of spare parts because of its ability to model the gradual degradation process of components and predict their condition. Also, according to previous studies, considering system information causes a significant reduction in costs. Therefore, the present study tries to extract the system information using a machine learning algorithm and provide it as a transition matrix to the Markov decision process (MDP) to determine the future states of the critical spare parts inventory system. In the presented method, the machine learning algorithm, here Adaptive Neuro-Fuzzy Inference System (ANFIS), is in charge of the training data. The Markov chain uses the trained data to predict the future states of the inventory system. For this purpose, four states have been considered, each representing a level of tension and demand in the inventory system. Applying the model to the data collected for a critical component showed that the model has good accuracy in predicting the following states of the system. Also, the presented model offers a lower error rate, RMSE, and MAPE, compared to the ARIMA model for predicting the next state of the inventory system