Staff scoring is an important aspect of workflow resource management. Staff scoring is the computational process for decision-making to identify the appropriate workforce with the skills required to perform a particular task. With staff scoring, the assignment of appropriate work to appropriate staff is critical for the success of tasks. However, efficiency is low due to manual processing by management personnel. The aim of this study is to develop staff scoring models. For this purpose, machine learning-based Random Forest (RF), Extreme Gradient Boosting (XGBoost) and Support Vector Machine (SVM) have been used with two different approaches to develop staff scoring models. Grid seacrh has been used to find the optimal values of the hyperparameters of the mentioned methods. One-hot encoding and frequency enconding have been used separately to convert some features in the dataset. The data is collected from the customers and staff of AlbertSolino, and includes 9 features and 557 rows. 10-fold cross-validation has been ustilized along with confusion matrix and accuracy have been used to assess the performance of staff scoring models. It has been observed that the developed staff scoring models show consistent performance and can be used for measuring which task is more suitable for the staff.
Anahtar Kelimeler: Staff Scoring, Machine Learning, Task Assignment
|
