Manufacturing of black tea is a complex process consisting of plucking, withering, curling, fermenting, drying, determining of quality, and packaging stages. Nowadays, tea quality is determined by two methods. Either chemical analysis method is used or tea tasters give quality scores by tasting, smelling, and visual inspection. By means of recent studies, tea quality can also be determined by electronic nose technology. The quality of tea is generally related with the fermentation stage. In this work, an electronic nose, which determines the quality of tea during the fermentation process in real time from only its odour, was made, and the fermentation band was controlled by this electronic nose. Fermentation bands of a tea factory in the Çayeli District of Rize, Turkey have been simulated and controlled via LabVIEW simulation software by using real odour values measured in the factory with the electronic nose. In the study, a new fermentation band system has been designed to function with the electronic nose. This section of the study has been performed in the simulation environment due to the high cost of physical system setup. With the method proposed in this study, tea factories will be able to increase their production capacity with higher quality.

Methanol, commonly used to cut costs in the production of counterfeit alcohol, is extremely harmful to human health, potentially leading to severe outcomes, including death. In this study, an electronic nose system was designed using 11 inexpensive gas sensors to detect the proportion of methanol in an alcohol mixture. A total of 168 odor samples were taken and analyzed from eight types of ethanol–methanol mixtures prepared at different concentrations. Only 4 features out of 264 were selected using the feature selection method based on feature importance. These four features were extracted from the data of MQ-3, MQ-4, and MQ-137 sensors, and the classification process was carried out using the data of these sensors. A Voting Classifier, an ensemble model, was used with Linear Discriminant Analysis, Support Vector Machines, and Extra Trees algorithms. The Voting Classifier achieved 85.88% classification accuracy before and 81.85% after feature selection. With its cost effectiveness, fast processing time, and practicality, the recommended system shows great potential for detecting methanol, which threatens human health in counterfeit drink production.

Agriculture is one of the most basic needs that is of vital importance for the human generation to be sustained. In agriculture, pesticides are an indispensable need both to increase productivity and to prevent crop losses. Some of these drugs are pesticides that protect the plant from harmful insects, fungi and rodents, while others are herbicides that protect it from weeds. However, if these chemicals penetrate into people as much as they ensure food safety, they can be harmful with various diseases in the long term. After these pesticides are applied to agricultural products, they decrease to safe levels after 3-10 days depending on the type and dosage of the drug. Long-term exposure to these chemicals can cause a number of health problems such as cancer, hormonal disorders, neurological diseases and immune system weakness. In this study, it was investigated whether the collected cherry and the pesticide-free cherry could be separated by smell after the pesticide was applied to it without passing the mentioned time. Cherries were collected from cherry trees that had never been sprayed, and then pesticide was sprayed on these trees, and cherries that were sprayed with pesticide were collected a day later. An electronic nose consisting of 11 very affordable gas sensors has been made for the study. One hundred pieces of odours were taken by the electronic nose, including different amounts of cherries with and without pesticides. Various attributes of these data have been extracted. Among the four different classification algorithms, the Extra Trees Classifier has given the most successful results with 94.30% classification accuracy, 93.00% sensitivity and 95.60% specificity classification success. The ability to detect the pesticides on the fruit with an electronic device is important for the monitoring of human health through food inspections.

Psoriasis is a chronic inflammatory skin disease with a high global prevalence. A skin biopsy is still required to diagnose the disease; no non-invasive diagnosis method has been found. It has become a popular approach for physicians as a support system, as it classifies biological data collected without human intervention in various ways with machine learning methods. Numerous studies have been conducted using machine learning methods to increase the accuracy, performance, speed, and reliability of diagnosing various diseases. This study aims to predict whether a group of patients admitted to Hitit University Erol Olçok Training and Research Hospital have psoriasis based on exhaled breath measurements using an electronic nose system which was produced for this study by the authors. In total, 263 clinical records were examined; 120 (45.6%) were obtained from healthy individuals, while 143 (54.4%) belonged to psoriasis patients. In order to distinguish data from those of psoriasis patients and those of healthy individuals, six different machine learning algorithms were used on the breath data set. The best classification result was provided by the ExtraTreesClassifier algorithm, with an accuracy rate of 96.1%, while other algorithms have rates between 66.6% and 94.2%. The most important outcome of this study is that the model determined to distinguish psoriasis patients from healthy ones can also help in the early diagnosis of psoriasis.

Power transformers are vital for transmitting, distributing, and using energy produced in electrical energy systems. The failure of a distribution transformer can cause significant financial losses to society and cause irreparable problems. Therefore, power transformers’ reliability, continuous monitoring, and fault-free operation are critical. In this study, an electronic nose system was developed to detect the duration of oil usage in power transformers based on its smell. In the system designed with eleven inexpensive gas sensors, a total of 200 transformer oil odors with four different usage periods were analyzed. Four features were selected from eighty-eight features using the Recursive Feature Elimination method with Grid Search and Cross-Validation, and they were classified with six different classifiers. With the Extra Trees algorithm, the most successful classifier, classification performance of 0.9810 CA, 0.9810 SE, 0.9937 SF was achieved without feature selection, and 0.9610 CA, 0.9610 SE, 0.9870 SF was achieved by selecting features. Dielectric breakdown voltage tests of the oils in the study were also performed, and the results supported the results of the electronic nose system. According to the results obtained, it can be concluded that transformer oil maintenance can be performed economically, practically, and reliably with the proposed system.