Salary Prediction in Data Science Field Using Specialized Skills and Job Benefits – A Literature Review

This paper aims to review the recent and existing methodologies for building a more suitable salary prediction model based on specialized skills and given job benefits in the Data Science field. The knowledge discovery also includes identifying existing human resource problems in the data science field and the most demanded skill set for early exploration and determination of input variables. As data science involves a high dimension of positions and responsibilities, the experimental dataset was projected to include skill-based and job benefits factors for more accurate salary predictions. The reviewed benchmarking machine learning methodologies on related problems are categorized into three main categories with individual strengths under different situations and requirements. Statistical methods are better in presenting variable relationships with extraordinary parameter tuning potential if linearity is present. Ensemble machine learning methods like Random Forest that combines multiple classifiers for more stable and accurate prediction. Deep learning-based neural networks have a strong specialty in handling unlabeled data and framework modifications. Moreover, it was realized that huge datasets with appropriate variables and grid search tuning method achieves greater and more reliable performance. However, extraordinary research on data science-related job benefits could be conducted if sufficient studies were present during that period. Overall, further work is necessary to determine the project’s objectives, scenario and experimental dataset to select suitable reviewed methodologies for the data science salary prediction model building.