Design And Implementation Of A Distributed Recruitment Management System

4.4    Comparative Evaluation
In this section we perform a comparison analysis between the results of my work against an existing system which utilises the same logical model; we also review the result of my work against the result of the various works reviewed in chapter 2.
4.4.1    Result of work VS Existing System
A method for carrying out exams without traditional face-face interaction between candidates and examiners is provided in (Garcia et al, 2010). The e-learning system described aims at supporting strategic decision concerning the evaluation of learners’ activities and can be used in adapting and customizing resource delivery (Zaiane et al, 2001); discovering and comparison with expected behavioural patterns specified by the instructor that describes an ideal learning path (Donnellan, 2003); giving an indication of how to best organize the educational web space and be able to make suggestions to learners who share similar characteristics (Bae, 2000); generating personalized activities to different groups of learners (Tseng, 2004); supporting the evaluation and validation of learning site designs (Machado, 2003); identifying interaction sequences indicative of problems and patterns that are markers of success (Maisonneuve, 2006).
It proposed a predictive association rule model aimed at the design of a global intelligent information system while integrating all decision support systems, process automation, all types of communication requirements and their interactions. Some of the main drawbacks of this model are that the used algorithms have too many parameters for somebody non expert in data mining and the obtained rules are far too many, most of them non-interesting and with low comprehensibility also due to the nature of the domain data size was usually small about (10-100) depending on the course (Hamalainen et al, 2006). Subsequently, we present another application of association rule mining to recruitment.
Another application of association rule mining for the determination of higher institutions which provide the best graduates in a particular field for the purpose of recruitment. The objective was to create an online recruitment test based system which creates a corporate memory consisting of candidate profiles for an enterprise to utilise when making decisions pertaining to recruitment. The model consists of four major steps: Data Gathering, Rule Definition, Rule Strengthening, and Result Selection.
The information management system modelled using unified modelling languages proposes different domain model based on the logical model to guide the domain expertise. The reasoning model uses a rule induction algorithm in which formal rules are extracted from a set of observations.
Just as their work includes the design of generic software architecture for a web-based Learning Management System, our approach also incorporates Decision Support architecture for web-based recruitment management system. However, our work integrates KDD technique for determination of schools in which company recruitment drive should be focused.
4.4.2    Result of work VS Reviewed work
The reviewed work was an Image retrieval system implementing fuzzy logic model and this was compared with the apriori algorithm implemented in the R.M.S. Fuzzy logic is a way of interfacing inherently analog processes that move through a continuous range of values, to a digital computer, that likes to see things as well-defined discrete numeric values. Simply put, fuzzy logic is the logic of approximate reasoning. A fuzzy engine is any automated machine whether hardware or software that employs the concept of fuzzy logic in its implementation pattern of decision making. The fuzzy engine is made up of three constituent parts or modules namely: the fuzzification module, the Fuzzy Rule Base/Evaluation Module, the defuzzification module. In the fuzzy rule base, fuzzified inputs, μ(x=A1), μ(x=A2), μ(y=B1), and μ(y=B2), are applied to the antecedents of the fuzzy rules. The fuzzy operator (AND or OR) is used to obtain a single number that represents the result of the antecedent evaluation. This number (the truth value) is then applied to the consequent membership function. To evaluate the disjunction of the rule antecedents, we apply the OR fuzzy operation. Typically, fuzzy expert systems make use of the classical fuzzy operation, union to represent disjunction: μ (A𝖴B(x)) = max [μ (A(x)), μ (B(x))] similarly, in order to evaluate the conjunction of the rule antecedents, we apply the AND fuzzy operation intersection: μ (A∩B(x)) = min [μ (A(x)), μ (B(x))]. Now the result of the antecedent evaluation can be applied to the membership function of the consequent.
In the case of association rule mining, an apriori algorithm is utilised where in users define minimum support and confidence that are tolerable, an implication rule is formed by formally describing data into premise and conclusion (A - > B). Interestingness measures are then added to strengthen the implication rule; these measures are in the form of confidence or support. The confidence classifier is derived by calculating the percentage of all transactions containing data item sets in both the premise and conclusion while the support is derived by calculating the number of transactions within the database containing data item sets in both the premise and conclusion, thus Association rule mining is formally defined as a process of finding the rules, where the support and confidence of the rule are greater than the user provided values of  minimum  support  and  minimum  confidence,  further  referred  to as minconf and minsup. The two values actually prune the search space and make mining possible.