Bayes ağları ile uluslararası rekabetçilik ölçümü
Measuring international competitiveness with Bayesian network
- Tez No: 884722
- Danışmanlar: PROF. DR. SEÇKİN POLAT
- Tez Türü: Yüksek Lisans
- Konular: Endüstri ve Endüstri Mühendisliği, Industrial and Industrial Engineering
- Anahtar Kelimeler: Belirtilmemiş.
- Yıl: 2024
- Dil: Türkçe
- Üniversite: İstanbul Teknik Üniversitesi
- Enstitü: Lisansüstü Eğitim Enstitüsü
- Ana Bilim Dalı: Endüstri Mühendisliği Ana Bilim Dalı
- Bilim Dalı: Endüstri Mühendisliği Bilim Dalı
- Sayfa Sayısı: 167
Özet
Bu tez, bir ülkenin bir sektörünün uluslararası rekabet gücünü ölçmek için bir model sunmaktadır. Porter'a göre, uluslararası rekabet gücü, yıllar içinde ekonomilerin yüksek bir performans düzeyine ulaşma mücadelesinde hedef haline gelmiştir. Uluslararası rekabet gücünü ölçen bu model, yüksek performans elde etmek için stratejiler geliştirmek ve bu geliştirilen stratejilerin rekabet gücü üzerindeki etkisini öngörmek için pratik faydalar sağlar. Bu çalışmada, Porter'ın uluslararası rekabet gücü çerçevesi (yani Porter'ın elmas modeli) doğrultusunda Bayes modeli oluşturulmuştur. Literatürde, bir ülkenin bir sektöründeki uluslararası rekabet gücü üzerine Porter çerçevesini kullanarak yapılan çalışmalar iki kısma ayrılabilir. Çalışmaların ilk kısmı doğası gereği nitelikseldir. Bu çalışmalarda, uluslararası rekabet gücünü etkileyen faktörler ve alt faktörler subjektif olarak değerlendirilmiştir. İkinci kısım çalışmalar ise faktörler arasındaki ilişkiyi analiz etmeyi amaçlayan doğrusal istatistiksel modellerden oluşmaktadır. İlk gruptaki çalışmaların ana sorunu subjektiflik, ikinci gruptaki çalışmaların ise değişkenler arasında doğrusal bir ilişkinin olmasıdır. Oluşturulan Bayes ağı modelinin bu sorunları aşacağı değerlendirilmektedir. Porter, uluslararası rekabet gücü çerçevesini elmas modeli aracılığıyla açıklamaktadır. Elmas modeli, birbirleriyle etkileşim halinde olan dört ana faktör (talep koşulları, faktör koşulları, firma stratejisi, yapısı ve rekabet ve ilgili ve destekleyici endüstriler) ile birlikte, tüm bu faktörleri etkileyen devlet ve şans (şans faktörünün tanımlanmasındaki zorluktan dolayı bu çalışmaya dahil edilmemiştir.) ile açıklar. Elmas modelinde ayrıca bu ana faktörlere ilişkin alt faktörler de bulunmaktadır. Uluslararası rekabet gücünün Bayes ağ modeli ile ölçülebilmesini sağlamak için değişkenlerin tanımları, ilişkileri, kategorileri ve olasılık değerleri belirlenmiştir. Bu alandaki sınırlı veriler nedeniyle, oluşturulan Bayes modeli veriye dayalı değildir. Geliştirilen model Porter'ın teorik değişkenlerini ve uzmanların olasılıklar konusundaki inançlarını birleştiren bir hibrit model sunmaktadır. Porter'ın çerçevesinde birçok alt faktör olduğundan, literatürden seçilen çalışmalardaki kullanım sıklığı nispeten yüksek olanlar Bayes modelinde kullanılmıştır. Uzmanların olasılıklar konusundaki inançlarını toplamak için anketler oluşturulmuştur. Geliştirilen Bayes ağ modeli GeNIe yazılımı aracılığıyla analiz edilmiştir. Modelin bütün olarak değerlendirilmesi ve sonuçların çeşitli koşullar altında analiz edilmesi için duyarlılık ve senaryo analizleri yapılmıştır. Modelin geçerliliği için, ilgili literatürdeki iki çalışmanın sonuçları, bu çalışmada geliştirilen modelle karşılaştırılmıştır. Sonuçlar, bu iki çalışmayla benzerdir. Ayrıca, duyarlılık analizlerinin sonuçları çoğunlukla Porter'ın çerçevesi ile uyumludur. Uluslararası rekabet gücünü en önemli derecede etkileyen faktörler ve alt faktörler analizlerle belirlenmiştir. Bu model, akademik amaçlar doğrultusunda uluslararası rekabet gücünün değerlendirilmesi için kapsamlı, niceliksel bir yaklaşım sunar. Bu çalışmanın sınırlaması, olasılıkların gerçek verilere dayanmaması, uzmanların inançlarına dayanmasıdır, bu nedenle model gerçekliği daha az temsil edebilir. Gelecekteki araştırmalarda, Porter'ın tanımladığı bir ulusun gelişim aşamalarına göre (faktör odaklı, yatırım odaklı ve inovasyon odaklı) Bayes modelde kullanılabilecek alt faktörlerin belirlenmesi, potansiyel bir çalışma konusudur.
Özet (Çeviri)
This thesis presents a model for measuring international competitiveness of a sector in a country. According to Porter, over the years international competitiveness has become a crucial goal for economies in their struggle for achieving a high level of performance. This model, which measures international competitiveness, provides practical benefits in developing strategies to achieve high performance and predicting the impact of these developed strategies on competitiveness. In this study, a Bayesian model is structured on Porter's international competitiveness framework (namely Porter's diamond model). In literature, studies on the international competitiveness of a sector in a country using Porter framework can be divided into two parts. First part of studies is qualitative in nature. In these studies, factors and sub-factors influencing international competitiveness have been subjectively evaluated. The second part of studies consists of linear statistical models that aim to analyse relationship among the factors. The main problem with the first type is subjectivity and the one with the second types is that they are based on a linear relationship between variables. It is evaluated that the Bayesian model will overcome these problems. Porter explains the framework of international competitiveness through his diamond model. The diamond model, which is the basis of the study, has been applied to more than 100 sample sectors or industries across 10 countries. In his book“Competitive Advantage of Nations,”Porter detailed with evidence how a nation can be more successful in an industry. The diamond model is explained by four main factors (demand condition, factor condition, firm strategy, structure and rivalry and related and supporting industries) that interact with each other, along with government and chance (not included in this study because of difficulty in defining) that influence all of these factors. The model also includes sub-factors related to these main factors. According to Porter's framework, the 4 main factors are explained as follows. •Factor conditions: The position of a nation in terms of the factors of production necessary to compete in a particular industry. (etc. skilled labor or infrastructure) •Demand conditions: The nature of domestic demand for the industry's products or services. •Related and supporting industries: The presence or absence of supplier industries and related industries that are internationally competitive at the national level. •Firm strategy, structure and rivalry: Conditions that determine how companies are established, regulated and managed within a nation, as well as the nature of internal competition. As far as we know, the Bayesian network approach for the Porter Diamond model is used for the first time in this study. To construct the Bayesian network model for international competitiveness, definitions, interrelationships, categories and probability values of the variables were determined. Due to limited data in this field, this Bayesian model is not data-based but rather a hybrid model integrating Porter's theoretical variables and experts' beliefs on probabilities. Since there are many sub-factors in Porter's framework, those with relatively high frequency of use in studies were selected from the literature were used in the Bayesian model. Through the literature review, we examined a total of 21 studies - 13 from Turkish sources and 8 from foreign sources - and the frequency of use of the sub-factors in these studies were determined. In addition, we carried out an analysis to determine whether there were sector and country clusters regarding the frequency of use of factors and sub-factors. The clustering seen most clearly and the least number of clusters was in the firm strategy, structure and rivalry. The highest number of clusters were occurred in factor and demand conditions. As a result, we observed that there was no clear clustering on the basis of sector and country according to sub-factor usage frequencies in the studies examined. By determining the main and sub-variables, the model needs to be made measurable. In this regard, categories that will make the model measurable for all factors were determined. We set category classifications with a maximum of 3 and a minimum of 2. The reason why the category classification were determined as 2 is that the number of questions increases exponentially when the number of categories increases due to the Bayesian network structure. In the categories, the lowest probability were determined as little, weak, low, or absent, in accordance with the factor and sub-factor meaning definitions, and the highest probability determined as very, strong, high, present. The main factors, sub-factors, and category classes for all factors were determined, and subsequently, the Bayesian network structure, which constitutes the main part of the thesis, were created. After that, questionnaires were developed to collect experts' beliefs on probabilities. An analysis of the belief provided by experts to the survey questions has been carried out. The probability values used in the analysis and the model were determined by calculating the arithmetic mean of the beliefs provided by the experts. Thus, we analyzed the standard deviations of the probability values answered by 11 experts. When the beliefs of experts about the probabilities were analyzed, it was determined that the most deviation among the probabilities occurred in the factor conditions, and the least deviation occurred in the related and supporting industries factor. The developed Bayesian network model was analysed through GeNIe software. The steps of using the model in the software were explained in detail. Sensitivity and scenario analysis have been carried out to evaluate the model as a whole and analyse the results under various conditions. For the validity of the model, we compared the results of two studies (Turkish electronics industry and Turkish plastic industry) in the related literature with the developed model in this study. The results are similar with these two studies. •According to the results obtained by applying the information found in the Turkish electronics sector competitiveness analysis study to the Bayesian network model, we observed that the results are the same in all factors except international competitiveness. •According to the results obtained by applying the information found in the Turkish plastics industry competitiveness analysis study to the Bayesian network model, we observed that the results are the same in firm strategy, structure and rivalry and demand conditions, but there are differences in the classifications of other factors. We also carried out sensitivity and scenario analyzes. It was considered important to show to what extent all sub-variables affect other variables. For this reason, sensitivity analyzes were carried out for all sub-factors and the importance rankings of all these sub-factors were explained. 10 scenarios have been developed to analyze the conditions of factors affecting international competitiveness under specific circumstances. Also, result of the sensitivity analyses and scenario analyses mostly comply with Porter's framework. Factors and sub-factors that most significantly influence international competitiveness have been identified regarding the analyses. Summary sensitivity analysis results, which one factor is changed and other factors are kept constant according to the detailed sensitivity analysis, are listed below. •The highest degree of weakness in international competitiveness occurred when the firm strategy, structure and rivalry were weak. This finding is consistent with Porter's statements that local competition is the factor that affects other factors the most. If competition is weak, international competitiveness will also be weak. •The highest level of strong in international competitiveness occurred when the development of factor conditions was strong. This finding shows that the effect of weak factor conditions on international competitiveness is lower than the effect of strong factor conditions on international competitiveness. This finding indicates that factor conditions make good conditions better. According to the detailed scenario analysis, the summary scenario analysis results, which are considered critical, are listed below: •In scenarios where one factor is selected as weak and the others as strong, we observed that the weak international competitiveness is mostly affected by firm strategy, structure and rivalry. •In scenarios where one factor is selected as strong and the others as weak, we observed that the strong of international competitiveness is mostly affected by firm strategy, structure and rivalry. Base on these findings, we can conclude that the most important impact on international competitiveness is firm strategy, structure and rivalry. This model provides a comprehensive, quantitative approach for academic purposes, providing as a structural foundation for assessing international competitiveness. The limitation of this study is that since the probabilities are based on experts' belief, not actual data, the model may be less representative of reality. Results may differ with more experts' belief. Porter stated in his study that the challenges that nations at different development levels may face may differ from each other. In addition, he stated that nations can make progress in terms of competitive advantage and competitive styles as they develop, and divided these stages into 3 (Factor oriented, investment oriented and innovation oriented). In the future research, determining sub-factors that may be used in Bayesian model according to the nation's development stages identified by Porter is a potential area.
Benzer Tezler
- İş sürekliliği yönetiminde analiz ve planlama için bayes ağlarına dayalı bir model
A model for analysis and planning in business continuity management with bayesian networks
ZEYNEP ÇAKIR SAYGILI
Yüksek Lisans
Türkçe
2019
Endüstri ve Endüstri Mühendisliğiİstanbul Teknik ÜniversitesiEndüstri Mühendisliği Ana Bilim Dalı
DOÇ. DR. UMUT ASAN
- Seyir emniyetinde insan hatası risk analizi ve insan faktörleri temelinde köprüüstü dizaynına yönelik kural önerileri
Human error risk analysis in navigational safety and human factors based rule recommendations on bridge design
RİFAT BURKAY ALAN
Doktora
Türkçe
2021
Denizcilikİstanbul Teknik ÜniversitesiGemi İnşaatı ve Gemi Makineleri Mühendisliği Ana Bilim Dalı
PROF. DR. OĞUZ SALİM SÖĞÜT
- Karşıtsal risk analizi ve melez bayes-kanaat ağları: Doğu Akdeniz problemi örneği
Adversarial risk analysis and bayes-belief networks: Eastern mediterranean problem example
GÜLŞAH BOZCU
Doktora
Türkçe
2024
Endüstri ve Endüstri Mühendisliğiİstanbul ÜniversitesiSayısal Yöntemler Ana Bilim Dalı
DOÇ. DR. ESMA NUR ÇİNİCİOĞLU
- Kimyasal tanker gemilerinde tank yıkama operasyonlarının entegrasyonu
Integration of tank cleaning operations on chemical tanker ships
TAYLAN GÜLER
Doktora
Türkçe
2024
Denizcilikİstanbul Teknik ÜniversitesiDeniz Ulaştırma Mühendisliği Ana Bilim Dalı
DOÇ. DR. İSMAİL ÇİÇEK
- Tedarik zinciri yönetiminde bilgi sistemleri: Havacılık yer hizmetlerinde uçuş zamanlaması için bilgi paylaşım modeli
Knowledge system utilisation in supply chain management: A model for knowledge sharing in aviation land services
MEHMET YÖRÜKOĞLU
Doktora
Türkçe
2013
Endüstri ve Endüstri MühendisliğiHava Harp Okulu KomutanlığıEndüstri Mühendisliği Ana Bilim Dalı
PROF. DR. SAMİ ERCAN
DOÇ. DR. GÜLGÜN KAYAKUTLU