Research Outputs
Permanent URI for this communityhttps://hdl.handle.net/20.500.14288/2
Browse
40 results
Search Results
Publication Metadata only A kernel-based multilayer perceptron framework to identify pathways related to cancer stages(Springer Science and Business Media Deutschland GmbH, 2023) Mokhtaridoost, Milad; N/A; Department of Industrial Engineering; Soleimanpoor, Marzieh; Gönen, Mehmet; PhD Student; Faculty Member; Department of Industrial Engineering; Graduate School of Sciences and Engineering; College of Engineering; N/A; 237468Standard machine learning algorithms have limited knowledge extraction capability in discriminating cancer stages based on genomic characterizations, due to the strongly correlated nature of high-dimensional genomic data. Moreover, activation of pathways plays a crucial role in the growth and progression of cancer from early-stage to late-stage. That is why we implemented a kernel-based neural network framework that integrates pathways and gene expression data using multiple kernels and discriminates early- and late-stages of cancers. Our goal is to identify the relevant molecular mechanisms of the biological processes which might be driving cancer progression. As the input of developed multilayer perceptron (MLP), we constructed kernel matrices on multiple views of expression profiles of primary tumors extracted from pathways. We used Hallmark and Pathway Interaction Database (PID) datasets to restrict the search area to interpretable solutions. We applied our algorithm to 12 cancer cohorts from the Cancer Genome Atlas (TCGA), including more than 5100 primary tumors. The results showed that our algorithm could extract meaningful and disease-specific mechanisms of cancers. We tested the predictive performance of our MLP algorithm and compared it against three existing classification algorithms, namely, random forests, support vector machines, and multiple kernel learning. Our MLP method obtained better or comparable predictive performance against these algorithms.Publication Open Access A model-based heuristic to the min max K-arc routing for connectivity problem(Schloss Dagstuhl - Leibniz-Zentrum für Informatik, 2014) Akbari, Vahid; Department of Industrial Engineering; Salman, Fatma Sibel; Faculty Member; Department of Industrial Engineering; College of Engineering; 178838We consider the post-disaster road clearing problem with the goal of restoring network connectivity in shortest time. Given a set of blocked edges in the road network, teams positioned at depot nodes are dispatched to open a subset of them that reconnects the network. After a team finishes working on an edge, others can traverse it. The problem is to find coordinated routes for the teams. We generate a feasible solution using a constructive heuristic algorithm after solving a relaxed mixed integer program. In almost 70 percent of the instances generated both randomly and from Istanbul data, the relaxation solution turned out to be feasible, i.e. optimal for the original problem.Publication Metadata only A multi-objective optimization approach for sustainable supply chains incorporating business strategy(IEEE, 2019) N/A; Department of Industrial Engineering; Bozgeyik, Esma Nur; Türkay, Metin; Master Student; Faculty Member; Department of Industrial Engineering; Graduate School of Sciences and Engineering; College of Engineering; N/A; 24956Sustainability is a necessity in the design and operation of supply chains. the triple bottom line (TBL) accounting of sustainability needs to incorporate economic, environmental and social pillars simultaneously in the decision making process. the business strategy can be developed to promote sustained growth, Also incorporating in the supply chain management issues as a business strategy rather than philanthropy. Deciding on the location of business facilities, supplier-manufacturer network, manufacturer-demand location network and the supplier- manufacturer relation strategy are among the important decisions in business strategy and supply chain management. However, there is a lack of theoretical work which analyzes the business strategy together with TBL concept of sustainability for the supply chain network design problem. in this paper, A methodological approach based on mathematical programming is proposed that conforms to the TBL accounting for supply chain network design problem from suppliers to customers embedded with business strategy and green energy usage option. a realistic case study is applied to the model. the results show that working with inclusive suppliers and using green energy are preferred with highest profit value.Publication Metadata only A multitask multiple kernel learning algorithm for survival analysis with application to cancer biology(JMLR-Journal Machine Learning Research, 2019) N/A; Department of Industrial Engineering; Department of Industrial Engineering; Dereli, Onur; Oğuz, Ceyda; Gönen, Mehmet; PhD Student; Faculty Member; Faculty Member; Department of Industrial Engineering; Graduate School of Sciences and Engineering; College of Engineering; College of Engineering; N/A; 6033; 237468Predictive performance of machine learning algorithms on related problems can be improved using multitask learning approaches. Rather than performing survival analysis on each data set to predict survival times of cancer patients, we developed a novel multitask approach based on multiple kernel learning (MKL). Our multitask MKL algorithm both works on multiple cancer data sets and integrates cancer-related pathways/gene sets into survival analysis. We tested our algorithm, which is named as Path2MSurv, on the Cancer Genome Atlas data sets analyzing gene expression profiles of 7,655 patients from 20 cancer types together with cancer-specific pathway/gene set collections. Path2MSury obtained better or comparable predictive performance when benchmarked against random survival forest, survival support vector machine, and single-task variant of our algorithm. Path2MSury has the ability to identify key pathways/gene sets in predicting survival times of patients from different cancer types.Publication Metadata only Admission policies for a two class loss system(Taylor and Francis, 2001) Burnetas, Apostolos; Van Der Wal, Jan; Department of Industrial Engineering; Örmeci, Lerzan; Faculty Member; Department of Industrial Engineering; College of Engineering; 32863We consider the problem of dynamic admission control in a Markovian loss queueing system with two classes of customers with different service rates and revenues. We show that under certain conditions, customers of one class, which we call a preferred class, are always admitted to the system. Moreover, the optimal policy is of threshold type, and we establish that the thresholds are monotone under very restrictive conditions. Copyright 2001 by Marcel Dekker, Inc.Publication Metadata only Aggregate planning problem from sustainability perspective(IEEE, 2015) Department of Industrial Engineering; Türkay, Metin; Saraçoğlu, Öztürk; Arslan, Mehmet Can; Faculty Member; PhD Student; PhD Student; Department of Industrial Engineering; College of Engineering; Graduate School of Sciences and Engineering; Graduate School of Sciences and Engineering; 24956; N/A; N/AAggregate planning refers to the determination of production, inventory and capacity levels for a medium term. Traditionally standard mathematical programming formulation is used to devise the aggregate plan so as to minimize the total cost of operations. However, this formulation is purely an economic model that does not include sustainability considerations. In this study, we revise the standard aggregate planning formulation to account for additional environmental and social criteria from sustainability perspective. We analyze the revised models and propose results that would be insightful for decision makers. We show how these additional criteria can be appended to traditional cost accounting in order to address sustainability in aggregate planning.Publication Open Access An improved lower bound on the competitive ratio of deterministic online algorithms for the multi-agent k-Canadian Traveler Problem(Finding Press, 2022) Shiri, Davood; Department of Industrial Engineering; Salman, Fatma Sibel; Faculty Member; Department of Industrial Engineering; College of Engineering; 178838We present an improved lower bound on the competitive ratio of deterministic online algorithms for the multi-agent k-Canadian Traveler Problem.Publication Metadata only An new approach to optimum component testing problem incorporating expected system lifetime(IFAC, 2009) Yamangil, Emre; Kuban Altinel I.; Feyziog̃lu, Orhan; Department of Industrial Engineering; Özekici, Süleyman; Faculty Member; Department of Industrial Engineering; College of Engineering; 32631We analyze the component testing problem of devices which consist of series connection of redundant, standby redundant and k-out-of-n subsystems. Although system reliability is a common performance measure, here we extend previous studies by considering expected system lifetime. This case applies when setting mission time for a system is more practical than deciding on system reliability accurately. The problem is formulated as a semi-infinite linear programming problem, and the optimum test times are obtained with a column generation technique incorporating reverse convex programming. The proposed solution technique is also illustrated by numerical examples.Publication Metadata only Analysis of mobile phone call data of İstanbul residents(IGI Global, 2015) Department of Industrial Engineering; Department of Industrial Engineering; N/A; Salman, Fatma Sibel; Sivaslıoğlu, Erbil; Memiş, Burak; Faculty Member; Undergraduate Student; PhD Student; Department of Industrial Engineering; College of Engineering; College of Engineering; Graduate School of Business; 178838; N/A; N/AIn this chapter, we analyze call detail records of subscribers of a major cellular network provider in Turkey with a focus on subscribers that reside in Istanbul. We consider a sample of 10,000 opt-in subscribers, chosen proportionally to the population density of each district of Istanbul. The anonymized cell phone usage data for 6 weeks are combined with demographic and subscription package attributes. Our methodology consists of data retrieval and cleaning, analysis and visualization. The analysis aims to extract information to be used mainly in disaster preparedness, marketing and public service design, and is categorized under: 1) understanding call habits in terms of call duration and call location with respect to gender and age categories, 2) tracking population density changes by time and district, 3) segmentation of people visiting specified locations, 4) information on mobility of disabled subscribers, and 5) international travel patterns by roaming data analysis.Publication Open Access AUC maximization in Bayesian hierarchical models(IOS Press, 2016) Department of Industrial Engineering; Gönen, Mehmet; Faculty Member; Department of Industrial Engineering; College of Engineering; 237468The area under the curve (AUC) measures such as the area under the receiver operating characteristics curve (AUROC) and the area under the precision-recall curve (AUPR) are known to be more appropriate than the error rate, especially, for imbalanced data sets. There are several algorithms to optimize AUC measures instead of minimizing the error rate. However, this idea has not been fully exploited in Bayesian hierarchical models owing to the difficulties in inference. Here, we formulate a general Bayesian inference framework, called Bayesian AUC Maximization (BAM), to integrate AUC maximization into Bayesian hierarchical models by borrowing the pairwise and listwise ranking ideas from the information retrieval literature. To showcase our BAM framework, we develop two Bayesian linear classifier variants for two ranking approaches and derive their variational inference procedures. We perform validation experiments on four biomedical data sets to demonstrate the better predictive performance of our framework over its error-minimizing counterpart in terms of average AUROC and AUPR values.