Papüloskuamöz deri hastalıkları halk arasında oldukça sık rastlanan ve kendine has morfolojik özellikleri olan deri hastalıkları grubudur. Papüloskuamöz deri hastalıklarının alt gruplarının belirtileri birbirine çok yakın olduğu için teşhis süreci bazı durumlarda zahmetlidir. Hastalığın teşhisi klinik muayenede konulabilir. Klinik muayenenin yetersiz olduğu durumlarda, tanı deri biyopsisi ile histopatolojik değerlendirme ile konulmaktadır. Bu süreçte dermatolog ve patoloğun uyumlu bir şekilde çalışması ve her iki hekimin de teşhis süreci ile ilgili bilgi birikiminin iyi olması gerekir. Bu yüzden Papüloskuamöz deri hastalıklarının tanısı deri biyopsisine ihtiyaç duyulmadan sadece klinik muayene ile dematolog tarafından konulabilmesi için daha basit, yüksek başarı oranına sahip ve klinikte kullanılabilir yöntemlere ihtiyaç duyulmaktadır. Bu çalışmanın amacı Papüloskuamöz deri hastalıklarının yüksek başarı oranı ile tespit edebilecek, klinikte dermatolog tarafından kullanılabilecek, yapay zeka yöntemleriyle geliştirilmiş kural tabanlı algoritma geliştirmektir. Çalışma kapsamında daha önce toplanmış veri seti kullanılmıştır. Veri setinde Papüloskuamöz deri hastalıklarının altı farklı alt grubu için klinik ve histopatolojik bulgular bulunmaktadır. Öncelikle veri seti ikişer sınıflı olacak şekilde gruplandırılmıştır. Daha sonra özellik seçme algoritmalarıyla klinik ve histopatolojik bulgular seçilmiştir. Daha sonra karar ağaçları yardımıyla kural tabanlı teşhis algoritmaları oluşturulmuştur. Çalışma sonucunda, sadece seçilmiş klinik bulgular kullanılarak ortalama %82.98 doğruluk oranı, 0.89 duyarlılık, 0.76 özgüllük oranıyla Papüloskuamöz deri hastalıkları kural tabanlı algoritmalar geliştirilmiştir. Sonuç olarak, bu çalışmada elde edilen sonuçlara göre, çalışma kapsamında geliştirilen algoritmalar, Papüloskuamöz deri hastalıklarının teşhisi için yapay zeka yöntemleriyle geliştirilen yüksek doğruluk oranına sahip kural tabanlı algoritmalar klinikte kullanılabilir.
Papulosquamous skin diseases are common skin diseases and have morphological features. The diagnosis process_x000D_
is sometimes troublesome, as the symptoms of the subgroups of papulosquamous skin diseases are very close to_x000D_
each other. The diagnosis of the disease can be made at the clinical examination. In cases where the clinical_x000D_
examination is insufficient, the diagnosis is made by histopathological evaluation by skin biopsy. In this process,_x000D_
dermatologists and pathologists should work in harmony, and both doctors should have a good knowledge of the_x000D_
diagnosis process. Therefore, more uncomplicated, higher success rate, and clinically practical methods are needed_x000D_
in order for Papulosquamous skin diseases to be established only by a clinical examination by a dermatologist_x000D_
without the need for a skin biopsy. This study aims to develop a rule-based algorithm that can detect_x000D_
Papulosquamous skin diseases with a high success rate, can be used by dermatologists in the clinic, developed_x000D_
with artificial intelligence methods. Within the scope of the study, the previously collected data set was used. The_x000D_
data set contains clinical and histopathological findings for six different subgroups of Papulosquamous skin_x000D_
diseases. Firstly, the data set is grouped into two classes. Then, clinical and histopathological findings were_x000D_
selected with feature selection algorithms. Then, rule-based diagnostic algorithms were created with the help of_x000D_
decision trees. As a result of the study, Papulosquamous skin diseases rule-based algorithms have been developed_x000D_
with an average of 82.98% accuracy rate, 0.89 sensitivity, and 0.76 specificity rate using only selected clinical_x000D_
findings. Consequently, according to the results obtained in this study, algorithms developed within the scope of_x000D_
the study, high-accuracy rule-based algorithms developed with artificial intelligence methods can be used in the_x000D_
clinic for the diagnosis of Papulosquamous skin diseases.
In job-shop production systems, orders are assigned to work centers according to their routes, and their operations are performed in this order. Production is becoming more and more complex with the increasing number of product lines and work centers with different routes. Decisions to be made according to the realtime monitoring of a dynamic production environment have become important. With the Fourth Industrial Revolution, information technologies are widely used in industries. A large amount of data is obtained from production tools that are capable of communicating with each other by means of Industry 4.0 and the internet of things. In this study, a simulation model of a production system that can collect data in real-time via sensors in work centers has been created and operation conditions have been determined. Then, work center / machine loading strategies were compared according to the delay periods of the jobs. The simulation model with the best loading strategy was run according to three different demand rates. Then data related with the delay status of the orders and the status of the work centers was obtained. The data were evaluated with data mining classification algorithms and rules were determined for delayed jobs. These rules were added to the simulation model as a decision mechanism. When an order is received in this model, the expert system estimates whether or not there will be a delay, and makes a decision to outsource the order’s production if needed. This approach further reduces the number of delayed orders
