A Competency Framework Model to Assess Success Pattern For Indian Faculties A NLP Based Data Mining Approach
R K Banu1* and R Ravanan2
*1Sathyabama University, Chennai,Tamil Nadu, India,
2Département of Statistics, Presidency College, Tamil Nadu, India,
Received 19 January 2016; revised 08 July 2016; accepted 20 September 2016
Faculties who help us grow as people are responsible for imparting some of life’s most important lessons. We learn through them, through their commitment to excellence and through their ability to make us realize our own personal growth.
The researchers look at the effectiveness by number of ways of assessing faculties. In our research work we analyzed and assessed the success pattern of college faculties based on Neuro-Linguistic Programming (NLP), a branch of Behavioral Psychology of the modern day. Using NLP Tools we pick up Behavior and Response Patterns in people in different life situations. The response patterns may vary in different contexts. Hence the patterns are checked in various contexts. The reports generated out of this assessment helps to identity their core competencies and the areas of improvement for their professional growth.
Keywords: Data mining, NLP, Sequence pattern mining, Prefix Span algorithm, Competency
Introduction
Data Mining may generate thousands of patterns A pattern is interesting if it is easily understood by humans, valid on new or test data with some degree of certainty, potentially useful, novel, or validates some hypothesis that a user seeks to confirm People develop different patterns that work well in certain contexts 1 These patterns determine their capabilities and skills-sets, attitudes and preferences, beliefs and values. Each pattern has its own merits and demerits..Data Mining is a process of sorting through large amounts of data and taking out only relevant information. The term Data Mining is often used to separate the process of discovery and prediction.
Forecasting and predicting models helps the prediction of future events.2 Neuro-Linguistic Programming (NLP) is the science of modeling the patterns of human behavior. NLP explores the inner workings of the human mind: how we think, how we develop our desires, goals and fears and how we motivate ourselves, make connections, and give meaning to our experiences.3Everyone has a mixture of strengths and preferences.It is known as VAK Analysis (Visual, Auditory & Kinesthetic). The questions are based on VAK 4. Sequences are
common, occurring in any metric space that facilitates either total or partial ordering 5 Sequence pattern mining is from given a set of sequences, we can find the complete set of frequent subsequence.
The sequential pattern mining problem was first introduced by Agrawal and Srikant in 6.Given a set of sequences, where each sequence consists of a list of elements and each element consists of a set of items, and given a user-specified min support threshold, sequential pattern mining is to find all of the frequent subsequences, i.e., the subsequences whose occurrence frequency in the set of sequences is no less than min support 7. The perception about teachers varies from student to student on their teaching quality. Students may relate teachers personality. Some students like the approach and others may not. 8 In our research work We are trying to prove that , the competency ratio is high, The faculty able to handle any kind of student and prepare them to meet the industry expectation once they complete their course. We identify the patterns exhibited by faculty and match it with the required competencies defined in Faculty assessment job role.
Individual report can be printed immediately after the test. . In current era, the subject knowledge alone would not help the Faculty in a current competitive environment. . the Skills required for the Faculty in the current trend is discussed in this paper which
—————————
*Author for Correspondence E-mail:karthiyainspire@gmail.com
helps the Faculty to identify their current strength and the reports generated out of this assessment deals with the areas of improvement to perform in a better way to attain international standard with best teaching practice and methods. The faculty details collected from various educational institutions. The assessment questions are based on NLP. The list of competencies and association rules set are discussed in Section 2.
The issues in the existing system , the methodology, algorithm used to measure the competency is discussed in Section 3. The results and discussions are in Section 4. Finally concluded in Section 5.
Competency Framework
The Faculty assessment problem is computationally, unwidely becoming huge and difficult to produce the performance evaluation with the increasing database in a size.9 Therefore , we propose a heuristic algorithm for finding the performance of the Faculty.
The aim of the algorithm is to reduce the larger database into smaller database and also minimize the cost. The following assumptions were defined in this framework by referring.10 The main role of the work is designing the competency assessment using NLP, There are 10 competencies are assumed. Each competency have 5 literals. The literals have 5 options are named as A,B,C,D,E . The assumed competencies are listed in (Table 1). Association Analysis aims at analyzing data to identify event occurrence Mining
association rules searching for interesting patterns among items in given data set.11 The Association rules are performed in two stages . One, The discovery of frequent set of items from the projected database.
Second one is generating the association rules from the item sets.12,13,14 The following association rules are framed, which gives Support Count as Threshold value These values are used to find the success rate of each competency.
A( X, “2”) ⇒ SP of Individual Competency ( “ A” ) B( X, “2”) ⇒ SP of Individual Competency (“B “) C( X, “2”) V ( A( X, “1” ) ^ B( X, “1”) ^ C(X,”1”) ^ D(X,”1”)) ⇒ SP of Individual
Competency (“C “)
D( X, “2”) ⇒ SP of Individual Competency (“D“) E( X, “2”) ⇒ SP of Individual Competency (“E“) Where, X is a literal and A,B,C,D, and E are candidates.
The selection of success pattern process done alphabetically. Using data mining technique the success rate of an individual is calculated.
Methodology
This section describes the working principles of the proposed algorithms with an illustration in order to determine success pattern and find excellence of the organizations. First the issues identified in the existing system.
Issues
Many tools are available to measure Faculties performance. Most of the research works are based on the direct questions such as Time management, Curriculum Planning , Student’s feedback about the Faculty, Students Result. Few tools are based on a concept called DISC (Dominance, Inducement, Submission and Compliance D- type individuals mostly task-oriented and good decision makers. They usually focus on results and the bottom-line. I- type individuals are outgoing and people-oriented. They tend to influence and inspire people. They usually focus on fun and talking. S- type individuals are reserved and people-oriented. They tend to support people. Focus on peace and harmony. C-type individuals are reserved and task-oriented. They tend to be more cautious. Focus on facts and rules or multiple intelligences (Rhythmic / musical, Visual, Verbal, Logical, Bodily kinesthetic, Interpersonal, Intrapersonal, existential, Naturalistic). These tools will measure limited patterns and will give a 30-40 pages generic report about a person. With that it is
Table 1—List of competencies Competency Name Description
Presentation skill Ability to present and persuade the students effectively.
Student Facilitation Ability to trust and support the belief of the students that they can do well.
Continuous learning Ability to gather relevant information to enhance knowledge and capabilities.
Curriculum planning
and Scheduling Ability to plan a sequence of learning experience, teaching methods and assessment criteria.
Research Skill Ability to systematically gather and interpret scientific data to gain new insights.
Creativity Ability to create new ideas and styles to awaken student creativity.
Positive Feedback Ability to monitor the performance of the students and give supportive feedback..
Work Coordination Ability to combine and integrate different elements for smooth functioning.
Democratic
Decision making Ability to involve other stake holders in decision making process.
Requirement
Analysis Ability to identify the needs, analyze the impact, understand fully the consequences and avoid conflicting issues.
difficult to match a person’s capability with the skill sets required in teaching. In our research work , the questions are based on NLP, With reference to ISCO standards10 the individual competencies are indented and measured in a Faculty, Each competencies have 5 attributes totally 50 attributes Identifying the strengths and the area of improvement which supports the professional growth and development of a teacher.
Helps the institution to retain competent teacher and protect students form incompetent teachers.
Heuristic Algorithm
In this Algorithm, initially construct the database based on the assessment. The assessments are taken from the different educational institutions. The data are collected for computing the success rate of the teaching faculty. Second, The association rules are formed in section 2.This is mainly used for finding the success rate of the given competency. The third step of the algorithm, The given database is scanned and taken the projected data by using prefixspan algorithm which is used to reduce the database size.
PrefixSpan algorithm is created for mining in projected databases. In this study our database is a long continuous sequence. In this heuristic algorithm is used PrefixSpan algorithm for projecting the database and to mine a continuous sequence database.
Significant issue of PrefixSpan algorithm’s main attribute is that PrefixSpan only grows longer sequential patterns from the shorter frequent this was our useful source of idea. One particularly important problem in the area of sequential rate mining is the problem of discovering all subsequences that appear on a given sequence database and have minimum support threshold that are defined in association rule. The difficulty is in figuring out what sequences to try and then efficiently finding out which of those are frequent15. Pattern- growth methods are a more recent approach to deal with sequential pattern mining problems. The key idea is to avoid the candidate generation step altogether, and to focus the search on a restricted portion of the initial database. PrefixSpan is the most promising of the pattern-growth methods and is based on recursively constructing the patterns, and simultaneously, restricting the search to projected databases16
Algorithm 1
Finding the competency of individual
Step 1. Construct the database based on assessment and add one empty column as success rate(SR).
Step 2. Set all association rules as threshold.
Step 3. The prefix span algorithm for getting the projected database .
3.1 Project all the demand prefixed with X
3.2 Scan the demands once and find frequent single candidate pathY that could be merged with X and generate a candidate cycle X’ as X+Y 3.3 Output X’ as a candidate cycle
3.4 For each X’, construct X’ projected demand S|X’
Step 4. Apply the association rule and set the competency in the projected database that is reflected in the main database also.
Step5. Repeat the process until there is no data available in the database.
The Algorithm 1 is repeated for all the persons in the organization and generates the new database which is contained number of person and success rate of the each competency. The second part of the algorithm is finding the success pattern of the each person which is used to provide the overall efficiency of educational institutions. But this is not applicable for only educational institution; it is also used in any organization for finding the performance of the staff.
Algorithm 2
Finding the success pattern ratio
Step 1.construct new database with ‘n ‘person and 10 competency from algorithm1.
Step 2.Replace success rate A as 10,B as 9,C as 8,D as 7,E as 6.
Step 3.Setup attributes weightage for competency.
(Table 2a)
Step 4.Calculate Success pattern ratio = SR∕k=110CW.
Step 5.Find the categories of each person based on assumption given in Table2a.
To achieve success pattern of an individual. The following assumptions are taken Such as A = 10, B = 9, C = 8 , D -= 7 and E =6 respectively for each person to arrive final table.The competencies identified in Table 1 are measured using algorithm1..
The success pattern ratio is calculated by given attribute weightage in table 2a. Using algorithm 2.
The individual category derived based on Table 2.b.
This Modified prefixspan algorithm(Finding the competency of individual) is mainly in the construction of projected database and multiple scans.The cost of constructing projected database can be minimized by creating pseudo-projections in
Table 2 (a)—Attribute Weight age of Competency
S. No Competency Name Attribute
Weightage
1 Presentation skill 4
2 Student Facilitation 3
3 Continuous learning 3
4 Curriculum planning and Scheduling 4
5 Research Skill 3
6 Creativity 4
7 Positive Feedback 4
8 Work Coordination 4
9 Democratic Decision making 3
10 Requirement Analysis 4
2 (b)—Individual Competency Category Success Pattern Ratio Category
91 -100 Over skilled
81-90 Skilled
71-80 Fit
61-70 Can Consider
Below 60 Must Improve
memory, so that they can be processed faster than the other mining methods and prefixspan algorithm also.
Results and Discussions
By applying Apriori Association rule, they made decision about how the teachers are efficient in conducting workshops, Seminars and conferences17. In our work. We analyzed the top competencies of a faculty who can fit for their job role as a teacher with Modified Prefixspan Algorithm with minimum cost and time. Which will help the management to decide whether they had a right choice in their recruitment process to fulfil the expectation of student community to meet the industry requirements. Suggestions also discussed to improve their skill to compete in current era. Using the algorithm 1 & 2. The final table 3 is arrived. The competency Graph obtained is given in (Fig.1). The process is repeated and done for almost 1000 faculties across various educational institutions in various discipline such as Arts, Science and
Table 3—Success Pattern for 20 Faculties
Faculty 1 2 3 4 5 6 7 8 9 10 Total SP SP Ratio Category
1 36 24 27 32 24 32 32 32 24 32 295 8.19 81.94 Skilled
2 36 21 18 36 24 36 36 32 24 32 295 8.19 81.94 Skilled
3 36 18 27 32 27 24 32 36 30 36 298 8.28 82.78 Skilled
4 36 27 27 36 24 32 32 32 27 36 309 8.58 85.83 Skilled
5 28 21 18 24 18 28 28 28 18 24 235 6.53 65.28 Can Consider
6 32 21 24 32 18 32 28 32 21 28 268 7.44 74.44 Fit
7 32 30 27 40 30 28 32 32 24 32 307 8.53 85.28 Skilled
8 40 27 27 32 30 28 32 36 30 32 314 8.72 87.22 Skilled
9 40 27 27 36 27 36 36 36 27 36 328 9.11 91.11 Over Skilled
10 40 18 21 28 24 28 40 28 18 36 281 7.81 78.06 Fit
11 40 27 30 40 30 32 40 40 24 36 339 9.42 94.17 Over Skilled
12 40 27 24 36 27 36 32 32 24 36 314 8.72 87.22 Skilled
13 28 24 21 32 24 32 32 32 24 32 281 7.81 78.06 Fit
14 36 18 21 28 21 24 24 24 18 28 242 6.72 67.22 Can Consider
15 32 27 27 36 30 32 36 40 24 36 320 8.89 88.89 Skilled
16 36 27 24 36 27 36 32 32 24 36 310 8.61 86.11 Skilled
17 40 24 21 32 24 32 32 32 24 32 293 8.14 81.39 Skilled
18 36 18 21 40 30 36 24 24 18 28 275 7.64 76.39 Fit
19 24 27 21 40 24 32 40 40 24 36 308 8.56 85.56 Skilled
20 36 27 24 36 30 36 32 32 24 36 313 8.69 86.94 Skilled
Fig.1—Competency Graph for 20 faculties
Engineering Faculties. Profile compatibility gives an understanding of the overall capability of an individual to deliver the expected performance in a specific job. The projected database once scanned for the competency, Next time, it will verify the remaining data from the main table. The process gets over in the Initial scan itself. The time and cost is reduced due to this. Our Modified Prefixspan Algorithm works efficiently.
Conclusion
Our research work helps the institution to identify and measure the individual competency and helps to keep right person in a right job. This is an eye opener for an individual faculty to identify the areas where they lack in their competency and scope for improvement. Once understand the reality then the faculty who is comes under the category, “Fit” can become skilled”, “Skilled” become “Over Skilled”.
References
1 Jiawei H & Micheline K, Data Mining: Concepts and Techniques, 1stedu , Morgan Kaufmann Publishers, (2000).
2 Aranuwa F O & Sellapan P, A data mining model for evaluation of instructors’performance in higher institutions of learning using machine learning algorithms, IJCCIT (2013) 2345 – 9808.
3 Tosey, P & Mathison J, Introducing Neuro-Linguistic Programming Centre for Management, Learning and development, School of Management, University
of Surrey,(2006). http://thepeakperformancecenter.com/
educationallearning/learning/preferences/learning-
4 Carl H M & Roddick JF, Sequential Pattern Mining – Approaches and Algorithms, Article in ACM (2013).
5 Srikant R & Agrawal R , Mining sequential patterns:
Generalizations and performance improvements, EDBT’,(1996).
6 Pal AK & Pal S, Evaluation of Teacher’s Performance:A Data Mining Approach, (2013) 359-369.
7 Brown N, What makes a good educator? The relevance of meta programmes, Assessment & Evaluation in Higher Education 29(5), (2004) 515-533.
8 Banu RK & Ravanan R, Implementation of Success Pattern Assessment (SPA) for Indian Teachers using data mining Technique, IJAER, (2014) 30199-30204 http://www.ilo.
org/public/english/bureau/stat/isco/ ISCO’s competency requirements for University and Higher Education Teaching profession.
9 Sukany M et al, DataMining:performance improvement in education sector using classification & clustering algorithm, ICCCE , (2012).
10 Gyorodi C & Gyorodi R 2002, Mining Association Rules in Large Databases, Proc of OEMES (2002) 45-50.
11 Gyorodi C & Gyorodi R, Holbhan S & Pater M 2002, Mining Knowledge in Relational Databasess, Proc of CONTI (2002) 1-6.
12 J D Holt & Chung S M, Efficient Mining of Association Rules in Text Databases CIKM (1999) 234-242.
13 Antunes C & Oliveira AL , Generalization of Pattern-Growth Methods for Sequential Pattern Mining with Gap Constraints, in Intl Conf Mac Learning and DM, (2003) 239-251.
14 Hnn J, Pei J & Yin Y , Mining Frequent Patterns without Candiate Gerneration, Proc of ACM-SIGMOD, (2000).
15 Nirmala G & Mallikarjuna PB, Faculty Performance Evaluation Using Data Mining, IJARCST (2014).
