A Comparative Study on Manual and Online Leave Management Systems

A Comparative Study on Manual and Online Leave Management Systems

Hastha, Mr. Sunith Kumar T

PG Student, St Joseph Engineering College, Mangalore Assistant Professor, St Joseph Engineering College, Mangalore

Abstract – Educational institutions are gradationally switching from traditional paper- grounded leave operations to online systems in the current digital period. The preferences of council tutoring and non-teaching staff when requesting leaves are examined in this review, with a special emphasis on whether they favour homemade processes or the ease of online platforms. Some workers still use homemade operations because they're strange with them or do not have access to digital tools, indeed though numerous workers value the speed and translucency of online systems. By differing the two styles, this study not only shows how effective and stoner-friendly each system is, but it also looks at how precisely each system captures and handles leave data. Institutions can produce further effective, inclusive leave operation systems that promote stoner comfort and functional effectiveness by taking these preferences into consideration.

Keywords: Staff, Leave Management System, Leave Request, Institution.

INTRODUCTION

A tool that makes it simple for workers and directors to request, authorize, cover, and manage leave is known as a hand leave operation system. workers at the maturity of institutions are entitled to a variety of leave options, including motherliness leave, exploration leave, study leave, sick leave, periodic leave, and leave without pay. These leaves are requested and proved in agreement with the institution's programs. Since it's in charge of managing diurnal operations and keeping track of labour force, the executive department is essential to any organisation. Historically, leave operations have been submitted manually. Under this system, tutoring staff generally draft a leave operation by hand and submit it to their Head of Department (HOD), who also sends it to advanced authorities for blessing or rejection. Although this approach has been used for numerous times, it's constantly clumsy, slow, prone to miscalculations, and challenging to maintain over time. Hence, the need for a robotic system that's briskly, error-free, with no paperwork, and easy to manage. As digital metamorphosis reshapes executive processes in educational institutions, the relinquishment of online leave operation systems is gaining instigation. These systems not only enhance functional effectiveness but also reduce paperwork, ameliorate translucency, and enable real- time shadowing of leave records.

still, the transition to digital systems is n't invariant across all staff members. Factors similar as digital knowledge, part designation, and access to technology influence whether a hand prefers homemade or online operation styles. While some non- teaching or aged staff members stick to traditional styles out of habit or tech discomfort, tutoring staff who are comfortable with digital tools tend to favour online systems. This emphasises how inclusive system design is essential. In order to produce adaptive leave systems that increase effectiveness and satisfaction, recent exploration recommends using machine literacy to examine stoner patterns, similar as department, designation, and background.

LITERATURE REVIEW

In educational institutions, managing hand leave has long been a pivotal but delicate executive task. In order to cut down on paperwork and expedite blessing procedures, Adamu (1) highlights the significance of having a structured leave operation system. His study describes the problems with homemade approaches, like tracking difficulties, detainments, and miscalculations, and suggests digital results to increase effectiveness and translucency. The wider function of Human Resource Information Systems (HRIS) in education is examined by Chugh (2). He emphasises how executive tasks, similar as tracking leaves, are being reshaped by digital platforms, which makes it simpler for organisations to manage labour force records efficiently. In a analogous tone, Fakeeh (3) explores the eventuality of Decision Support Systems (DSS) for leave- related planning and soothsaying in advanced education and how they can enhance executive decision- timber. A case study on the development and deployment of an electronic leave operation system at a university is handed by Adisa (4). The system enhanced responsibility and record- keeping while streamlining the leave request and blessing procedure. These advancements follow a growing trend in which traditional paper- grounded workflows are being precipitously replaced by digital tools. Recent exploration has concentrated on incorporating machine literacy to advance beyond robotization into intelligent decision- timber. Ivanoti et al (5)(6) developed a decision support system for hand leave vaticination using

LightGBM and K- Means clustering. According to their findings, machine literacy models can prognosticate leave geste with high delicacy, which helps HR departments plan their pool. also, in order to comprehend hand geste, including waste pitfalls, Krishnamoorthy et al (7) and Gandhi et al (8) employed prophetic analytics in HR. These studies punctuate the significance of data- driven decision- making by showing that hand data, including department, performance, and engagement, can be used to make prognostications. By using tree- grounded classifiers to model hand waste, Cherian et al (8) expanded on this strategy and demonstrated the efficacity of ensemble literacy ways like XGBoost and Random timbers in analysing staff trends. This strategy fits in nicely with the current study's vaticination of leave operation preferences. By soothsaying stress situations using hand engagement and performance criteria, Latha et al (9) advanced prophetic modelling. Their exploration lends credence to the notion that advanced analytics can be applied to enhance institutional results and hand well- being in addition to geste vaticination. All effects considered, these studies support the transition of educational institutions from homemade to digital systems and demonstrate how machine literacy can be used to read and comprehend hand geste Combining intelligent algorithms with structured digital systems offers a promising path forward as organisations strive for staff satisfaction and functional effectiveness.

prejudiced towards the maturity class. In this study, three supervised machine literacy algorithms were used XGBoost Classifier, Random Forest Classifier, and Logistic Retrogression. These models were chosen because they can handle both direct and nonlinear connections and have a track record of success in bracket tasks. Each model was trained and tested singly after the dataset was divided into training and testing subsets in an 8020 rate. We employed bracket criteria similar as delicacy, perfection, recall, and F1- score to assess the model's performance. To make it apparent how numerous prognostications were accurate and where the model erred, we also constructed confusion matrices. For case, we were suitable to determine the frequence with which the model rightly or inaptly prognosticated a person's leave operation system. These criteria handed us with a clear and comprehensive picture of how well each algorithm could anticipate, given a staff member's profile and background, whether they would prefer to apply for leave online or manually. Incipiently, to determine which variables had the biggest goods on vaticination results, point significance was examined, particularly in the Random Forest and XGBoost models. Drawing perceptive conclusions about hand geste and preferences and offering helpful suggestions for creating further inclusive and effective leave operation systems in educational institutions depends heavily on this interpretability step.

FLOWCHART

METHODOLOGY

The purpose of this study's methodology is to ascertain whether tutoring and non-teaching workers in educational institutions prefer to apply for leave online or manually. A methodical, data- centric methodology that includes dataset creation, preprocessing, model structure, and performance evaluation is the foundation of the study. A synthetic dataset comprising 500 records was created in order to replicate authentic institutional data. Features like staff designation, department, gender, leave type and duration, leave status, access to digital tools, and the system of leave operation are all included in each record that's material to leave operation opinions. In order to separate between academic (similar as professors and adjunct professors) and executive (similar as clerks and librarians), a deduced point called isTeachingStaff was also introduced grounded on staff designation. To get the dataset ready for machine literacy operations, preprocessing was done after data generation. This included using one-hot encoding to render categorical variables, handling missing values, and spanning numerical features as demanded. After that, the data was divided into features (X) and target markers (y), with operation system acting as the variable that demanded to be prognosticated. The training set was subordinated to SMOTE (Synthetic Minority Over-sampling fashion) in order to amend any possible imbalance in the dataset between the primer and online operation classes. This saved the models' strong generalisability and averted them from getting

The methodological procedures used to read hand preferences for

leave operation styles are depicted in the flowchart. After data collection, preprocessing ways like point engineering and garbling are performed. To construct and assess the models, the dataset was also divided into training and testing sets.

Figure 1: Methodology Flowchart

The effectiveness of colorful machine learning algorithms, including Logistic Retrogression, Random Forest, and XGBoost, in prognosticating staff leave operation preferences was estimated using performance criteria, including delicacy, perfection, recall, and F1- score.

RESULT

Figure 2: Final Result

Staff preferences for online versus homemade leave operation processes were successfully classified by the XGBoost model with a high prophetic delicacy of 87.0. In the confusion matrix, class 0 (58 total samples and 51 correct prognostications) represents the online system, whereas class 1 (42 aggregate samples and 36 correct prognostications) represents the homemade system. This suggests that the maturity of workers prefer to apply for leave online. The raised support for class 0 and the enhanced perfection and recall support this trend, indicating that tutoring staff members, particularly those oriented to digital tools, are decreasingly using online systems due to their speed and stoner- benevolence. nevertheless, the model also reveals that a significant portion of workers continue to use the homemade system, which reflects varying degrees of technological comfort among colourful staff positions.

DISCUSSION

The study's findings show a distinct trend towards online leave operation processes, especially among tutoring staff. Using characteristics like designation, leave type, and digital access, the XGBoost classifier demonstrated a strong capability to separate between workers who prefer homemade or online systems, achieving a delicacy of 87. The online order reckoned for the maturity of accurate prognostications, suggesting that digital platforms are extensively accepted for submitting leave requests. Academic staff members' adding comfort and familiarity with technology is reflected in this trend. The speed, ease, and effectiveness of online systems naturally appeal to tutoring faculty, who are constantly more habituated to using digital tools for communication and instruction. A one- size- fits- all strategy might not work, however, as some non-teaching staff still prefer homemade operations, most likely out of habit or a lack of digital knowledge. As a result, organisations need to balance promoting

digital relinquishment with making sure that all workers can use it. Overall satisfaction can be increased by offering training, enhancing system armature, and continuing to support both online and homemade processes during transitional ages. Chancing patterns and allocating coffers meetly can be backed by the operation of machine literacy models similar as XGBoost.

CONCLUSION

This study reveals a growing preference for online leave operation systems over homemade bones, pressing a significant shift in the leave operation process among council staff, especially tutoring faculty. Systems that give speed, convenience, and translucency are easily preferred as educational institutions embrace digital metamorphosis more and more. The study reveals that the maturity of workers, particularly those who are more habituated to using digital tools and platforms, choose online approaches because of their availability and expedited workflow. The study successfully prognosticated staff preferences grounded on a number of characteristics, including department, designation, and access to digital structure, by utilising machine literacy models, particularly the XGBoost classifier. The model's high delicacy of 87 shows how prophetic analytics can be used in executive settings. These compliances can help organisations in creating further intelligent, stoner- concentrated leave operation programs that ameliorate hand satisfaction while contemporaneously satisfying functional conditions. But the study also discovered that a sizable portion of workers continue to use homemade operation procedures. This implies that inclusivity is still pivotal indeed though digital systems are getting more and more popular. Staff geste may be told by rudiments like a lack of access to technology, resistance to change, or low situations of digital knowledge. thus, in order to ensure that no group is left before in the digital shift, institutions must strive to develop mongrel or adaptable leave operation systems that support both homemade and online preferences. By automating repetitious tasks and offering practicable perceptivity that support executive departments in making data- driven opinions, prophetic analytics integration into similar systems can have a number of benefits. In the end, these developments may affect in better policy expression, more effective HR procedures, and increased hand engagement throughout the organisation.

REFERENCES

1. Adamu, Abubakar. EMPLOYEE LEAVE MANAGEMENT SYSTEM. FUDMA JOURNAL OF SCIENCES, vol. 4, no. 2, Jul. 2020, pp. 8691. DOI.org (Crossref), https://doi.org/10.33003/fjs-2020-0402-162.

2. Chugh, R. (2014). Role of Human Resource Information System in Educational Organization. Journal of Advanced Management Science.

3. Fakeeh, K. A. (2015). Decision Support System (DSS) in Higher Education System. International Journal of Applied Information System (IJAIS), 9(2).

4. Samuel Mayowa, Alade, et al. Design and Implementation of a Web Based Leave Management System. International Journal of Computer

   Applications Technology and Research, vol. 11, no. 04, Apr. 2022, pp. 123

   44. DOI.org (Crossref), https://doi.org/10.7753/IJCATR1104.1006.

5. Garg, Umang, et al. Classification and Prediction of Employee Attrition Rate Using Machine Learning Classifiers. 2024 International Conference on Inventive Computation Technologies (ICICT), IEEE, 2024, pp. 60813. DOI.org (Crossref), https://doi.org/10.1109/ICICT60155.2024.10544966.

6. Cuarez, Ryan O., et al. Streamlining Human Resource Leave Management System. 2024 Global Conference on Comunications and Information Technologies (GCCIT), IEEE, 2024, pp. 16. DOI.org (Crossref), https://doi.org/10.1109/GCCIT63234.2024.10862298.

7. Krishnamoorthy, N., et al. HR Analytics and Employee Attrition Prediction Using Machine Learning: Advances in Computational Intelligence and Robotics, edited by Jingyuan Zhao et al., IGI Global, 2024, pp. 7996. DOI.org (Crossref), https://doi.org/10.4018/979-8-3693-0683-3.ch004.

8. Gandhi, Anju Bhandari, et al. Employee Attrition Factors Based on Data Analytics. 2022 IEEE International Conference on Current Development in Engineering and Technology (CCET), 2022, pp. 16. IEEE Xplore, https://doi.org/10.1109/CCET56606.2022.10080699.

9. Cherian, Jerly Akku, et al. Predictive Modeling of Employee Attrition Using Tree-Based Machine Learning Classifiers. 2025 Emerging Technologies for Intelligent Systems (ETIS), IEEE, 2025, pp. 16. DOI.org (Crossref), https://doi.org/10.1109/ETIS64005.2025.10961183.

10. Latha, M. Pusha, et al. EarlyAlert: Predicting Employee Stress Through Performance and Engagement Metrics. International Journal of Recent Advances in Engineering and Technology, vol. 14, no. 1, Apr. 2025, pp. 10615. journals.mriindia.com,

    https://journals.mriindia.com/index.php/ijraet/article/view/183.

11. Iswarya, M., et al. Leave Management Portal. 2022 1st International Conference on Computational Science and Technology (ICCST), IEEE, 2022, pp. 21114. DOI.org (Crossref), https://doi.org/10.1109/ICCST55948.2022.10040328.