Prediction with machine learning algorithms for computer engineering department and evaluation with explainable artificial intelligence

This thesis study aims to predict the professional fields in which individuals in computer engineering can work, based on their technical competencies and academic performance, through machine learning. A custom dataset comprising 174 records was constructed, including variables such as grade point average (GPA), internship domain, project experience, database knowledge, and proficiency levels in programming languages such as Python, Java, C#, and C++. During the data preprocessing phase, categorical variables were encoded, and the dataset was prepared for modeling. For model selection, multiple classification algorithms were compared using LazyClassifier, and the top eight models exhibiting the highest accuracy were selected for further analysis. These models were trained on a shared dataset with hyperparameter optimization performed via GridSearchCV. Performance was evaluated using metrics including accuracy, precision, recall, and F1-score. The results demonstrated that the Gradient Boosting algorithm achieved the highest accuracy of 97%. To interpret the model's decision-making process, SHAP (SHapley Additive exPlanations) and LIME (Local Interpretable Model-agnostic Explanations) algorithms were employed. SHAP analyses highlighted the influence of specific projects and relevant domains on class-specific decisions, while LIME elucidated the rationale behind decisions related to database and programming language proficiencies. The findings indicate that the predictions are not only accurate but also interpretable. The developed model holds potential for effective application in career guidance, personnel placement, and educational planning.