1582 Ijiet 2959

Download as pdf or txt
Download as pdf or txt
You are on page 1of 9

International Journal of Information and Education Technology, Vol. 12, No.

1, January 2022

Evaluating Pangasinan State University Faculty


Performance Using Associative Rule Analysis
Frederick F. Patacsil, Paulo V. Cenas, Bobby F. Roaring, Jennifer M. Parrone, and Daniel Bezalel A.
Garcia

 (SEF).
Abstract—Evaluating faculty members' performance is a In the Philippines, teacher evaluation is part of the
very complex area to study. In addition, predicting the educational institution's practice. The purpose of student
performance of these faculty members is a very difficult and evaluation of faculty is to facilitate faculty growth and
challenging task. However, the core of education is teaching and
learning, and teaching-learning works to its fullest when there
development. Likewise, student evaluation is used to play a
are effective teachers. Measuring the effectiveness of faculty support tenure, promotion, transfer, and termination
members is done based on the student evaluation of faculty. decisions as well as salary increases and faculty awards. In
This research aims to develop a model to predict the Vietnam, improving the quality of teaching performance,
performance of the faculty members using associative rule professional enhancement, and making positive
based on the existing evaluation form used by PSU to evaluate improvements among teachers is the main goal of teacher
faculty members. The model is designed to utilize the
knowledge of text analytics rule capabilities that will provide
evaluation [2], [3]. The main objective of educational
great support for the decision-making of Pangasinan State institutions is to provide a quality education through
University in the Philippines. The result reveals that the term regular teaching assessment to maintain quality of
good is still the top one terms occurred for all campuses education [4]. According to Magno [5], this criterion implies
followed by teaching. The results indicated that teacher/faculty that the educational institution measures the quality of the
members on all campuses are good teachers. Associating words services through the delivery of instruction to their students.
reveal that "teaching good subject/topic," "explains simply"
Teaching is a development process that improves faculty
and other meaningful associated words can be utilized to
evaluate the performance of the teacher. The results exposed performance or a reviewing process that assists
not only the quantitative values of faculty evaluation it also administrators in making personnel and strategic decisions
exposed the qualitative opinion of the students in the [6]. The main objective of faculty evaluation is to identify the
performance of their faculty members. This study reveals strengths and work on the weak areas to improve the
important aspects of the faculty member's teaching performance of the teachers, thus achieve optimum process
performance in terms of words/association of words that will
quality. According to Sanders & Horn [7], Once teachers are
describe their teaching performance. The results can be utilized
in coaching and mentoring faculty members to cope with their given feedback about classroom-level instructional outcomes,
weaknesses. The proposed model can be utilized by Pangasinan they start to modify their instruction to address their weak
State University to evaluate the faculty members in terms of areas.
their teaching performance by utilizing the comments/opinions Evaluating teachers' performance is a difficult task, but it
of the students. is made for years through pedagogical surveys. The most
commonly used tool to evaluate a teacher's performance is
Index Terms—Association rule, data analytics, faculty
through surveying students' responses. The concern about
performance.
student evaluations is based mainly on 1) students not having
enough experience and maturity to evaluate the course and
I. INTRODUCTION the teachers, 2) students' evaluation is affected by the
popularity of the course and/or teacher, grades are given by
The core of education is teaching and learning, and the teacher and course being compulsory: or elective.
teaching-learning works to its fullest when there are effective Furthermore, the teacher may have the best performance on
teachers. The effectiveness of teaching can be measured research to find out the solution for community problems,
through the use of student evaluation of faculty, a survey that have good performance in community service, social life
students complete each academic semester. It is used to inside and outside the university [8]. To evaluate teacher's
determine the perception and knowledge of faculty members performance, students individually fill a pedagogical survey,
on the different basic and fundamentals of teaching, research, and the quantitative mean of the answers corresponds to the
and service using the available resources [1]. This evaluative teacher evaluation. However, despite the spread of such tools,
procedure is commonly referred to as student ratings of there is no study on automatically predicting teachers'
teaching (SRT), teacher rating form (TRF), student performance using qualitative surveys that will directly
evaluation of teaching (SET), or student evaluation of faculty identify the strength and weaknesses of a teacher [3].
Nowadays, data mining is becoming a more common
Manuscript received June 9, 2021; revised July 28, 2021. platform in understanding and solving educational and
The authors are with Pangasinan State University, Philippines (e-mail: administrative problems in higher education. Data mining
[email protected], [email protected],
techniques are applied in higher education more and more to
[email protected], [email protected],
[email protected]). give insights into educational and administrative problems to

doi: 10.18178/ijiet.2022.12.1.1582 21
International Journal of Information and Education Technology, Vol. 12, No. 1, January 2022

increase managerial effectiveness. For example, it is used to included. Refer to Table I for the sample comments derived
investigate factors related with the assessment of teaching from the actual data set. The estimated number of data
performance of educators [9]. However, most of the gathered is 15,548.
researches carried out on educational mining focus mainly on
TABLE I: SAMPLE COMMENTS FROM ACTUAL DATA SET WITH NUMBER OF
modeling student's performance, but there is very a small LANGUAGE OR DIALECTS USED
number of research model on faculty performance [10]-[13]. Number of
Data mining is a powerful technology for analyzing Language Comments
important information from data warehouses. It is a data Used
She is good at teaching. she do everything just to
analysis methodology used to identify hidden patterns in a understand all the topics that she discuss. Sometimes she is
large set of data [2]. With data mining techniques, such One (English) late because of first subject she handle but we understand
knowledge can be extracted and accessed in transforming the that because she wants us/them learned more about the
topics she gave
database tasks from storing and retrieval of learning and
One (Filipino,
extracting knowledge. It consists of a set of techniques that specifically Tamad magturo,madalang.
are used to extract relevant and interesting knowledge Tagalog)
from data. Data mining has several tasks such as Matalino si Sir. Time is gold para sa kanya, ayaw niya ng
Two (Filipino plagiarism at higit sa lahat may sariling pinaniniwalaan
association rule mining, classification and prediction, – English) Nag e-alien talk di ko na maintindihan minsan sa sobrang
and clustering. Classification techniques are supervised bilis o lalim ng mga words na ginagamit niya
learning techniques that classify the data item into the
predefined class label. It is made of the most useful Data translation was then employed after the data were
techniques in data mining to build classification models from compiled. Comments that are written in Filipino or
an input data set. It is the process of finding a set of models Tagalog-English were translated to the English language.
that describe and differentiate data classes and concepts, to Filling in the missing values from the comments such as
predict the class whose label is unknown [9]. The users missing letter/s were next employed. Care was taken to make
commonly build models used to predict future data trends. sure that the translation is consistent throughout the dataset.
There are several algorithms for data classification such as For example, comments like "instructing being not good in
decision tree (j48), Support Vector Machines (SVM), and teaching" is translated as "bad in teaching." Lastly,
Naive Bayes classifiers. With classification, the inconsistencies such as misspelled words or grammatical
generated model would be able to predict a class for errors were then corrected. The aforementioned processes
giving data depending on previously learned information were accomplished by the data processor.
from historical data [4]. B. Data Pre-processing
This research aims to develop a model to predict the
Data pre-processing was immediately employed after
performance of the faculty using associative rule based on the
cleaning the data gathered. Data pre-processing operators
existing evaluation form use by PSU to evaluate faculty
include tokenization, filtering stop words, stemming, case
members.
transformation, and feature selection or n-gram. Translation,
Specifically, the study aims to:
tokenization, case transformation, and filtering stop words
1) Determine the most frequent word/words used to describe
are the only data preprocessing steps that were employed.
the performance of a faculty member.
The Fig. 1 shows the data preprocessing techniques used in
2) Determine the performance of faculty members of every
this paper.
campus of Pangasinan State University using the
associative approach text analytics technique.
The proposed model is designed to utilize the knowledge
of associative text analytics capabilities that will provide
great support for the decision-making of Pangasinan State
University in the Philippines. Fig. 1. The preprocessing of data in RapidMiner9.7.

Tokenization is the first process employed under the data


II. MATERIALS AND METHODS preprocessing procedure. This operator breaks up the text of
a document into pieces of words, keywords, symbols, and
A. Data Collection other similar things called tokens. Several options can be
The first step in this paper is to gather all the necessary data. used in breaking up the text, but the default setting which is
Comments from the evaluation rating of the teaching staff of the non-letters was employed in this paper.
Pangasinan State University comprise the data for this study. Transform cases were then employed after the data set was
Data were collected from the PSU Online Portal. tokenized. This operation transformed the cases of letters
Questionnaires were sent to each students' portal so that they from lower case to upper case or vice versa. In this paper, all
could evaluate the performance of their instructors. Name of letters in the data were transformed into the lower case for
a faculty member, subject taught that semester (First convenience and at the same time avoiding mix-ups.
Semester, A.Y. 2019 - 2020), and comments were gathered. The final preprocessing step employed is the Filter
Most of the comments are in the English language. Stopwords (Dictionary) operator. This is done to remove all
Comments that are written in Filipino or those comments the tokens equal to the stopwords from the given text file and
with two or more languages or dialects used were also can be accomplished by listing down words that are assumed

22
International Journal of Information and Education Technology, Vol. 12, No. 1, January 2022

to be unnecessary. Another option is to download a text file II.


or document file containing pre-identified English stopwords
C. Numeric to Binomial
via the Internet. The Fig. 2 shows the sample saved text file
used in the paper to filter stopwords. If the document is not in The Numeric to Binomial operator was utilized after the
text file format, it is converted into that format because it is data preprocessing. This operator transforms the type of any
the must-be-done requirement of RapidMiner9.7. The latter numeric attribute into a binary, whether true or false. This
option was employed; the content was verified just in case process is necessary because the output after utilizing this is
that there are words that must be included or removed. operator is then connected to FP-Growth to determine
Further, adjectives and some pronouns (such as he and she) frequent itemset. Values for minimum and maximum
were not included in the list of stopwords. parameters must be specified first. If the value of an attribute
is between the specified values for the minimum and
maximum parameters, then it will be mapped to false.
Otherwise, the attribute will be mapped to true.
In this research paper, the default values for the minimum
and maximum parameters were used, that is, both are set to 0.
This means that if the value of an attribute is equal to 0, then
it will be mapped to false. All values, except for 0, are
mapped to true.
D. Vector Creation
The schema used in vector creation was TF-IDF or Term
Frequency – Inverse Document Frequency. This schema is a
numerical statistic that shows the importance of a word is to a
document in a corpus and considered the most popular
weight schema. In this schema, the TF-IDF value is directly
proportional to the incidence of documents in the corpus that
contain the word or term. Hence, as the word or term occurs
most often, the value increases.
The TF-IDF will choose those words that weighted greater
importance based on the dataset. The TF-IDF value increases
proportionally to the number of times a term appears in the
document but is offset by the frequency of the word in the
corpus. It helps to control or adjust because some terms are
Fig. 2. Saved text file of stopwords. generally more common than others.
TABLE II: THE BEFORE AND AFTER USING TRANSLATION AND DATA E. Pattern Recognition
PREPROCESSING
Process Before After The method that was applied in determining the pattern
She has good teaching She has good teaching recognition of comments was the Association Rule Mining.
skills. Nauunawaan ko skills. I understand In Association Rule Mining, all itemsets must meet the set
ang mga tinuturo niya. what she teaches.
Translation Minsan busy siya kaya Sometimes busy that's
value for the minimum threshold for support and confidence
hindi kami nakakapag why we can't meet but to arrive at a strong relationship between or among items.
meet paero ok lang it's okay. The formula for computing the support and confidence is
naman.
She has good teaching she has good teaching
given below:
skills. I understand skills. i understand
what she teaches. what she teaches. SUPPORT = occurrences of [A, B] /total number of transactions (1)
Transform Cases
Sometimes busy that's sometimes busy that's CONFIDENCE [A,B] = occurrences of [A,B] /total number of [A] (2)
why we can't meet but why we can't meet but
it's okay. it's okay.
she has good teaching she has good teaching Frequent itemsets are quantified by support which is
skills. i understand skills i understand defined as the ratio between the number of instances where
what she teaches. what she teaches [A, B] appeared together in a single transaction and the total
Tokenization
sometimes busy that's sometimes busy thats
why we can't meet but why we can't meet but
number of transactions. On the other hand, confidence is the
it's okay. its okay probability of finding [item A, item B] together.
she has good teaching she has good teaching This paper used a two-step process in analyzing the
skills i understand skills understand she
comments. The first step includes conducting comment word
Stopwords what she teaches teaches busy can't meet
(Dictionary) sometimes busy thats okay analysis using FP – Growth in terms of frequent patterns of
why we cant meet but itemsets. The value set for the minimum frequency for
its okay overall data (i.e., compiled data of every campus) is different
for the per-campus data. The value set for the overall data is
The tokenization, case transformation, and filtering stop 67 and per campus differs depending on the data set. The
words were done using the RapidMiner. The results after second step includes the examination and determination of
applying data preprocessing techniques are shown in Table the strong relationship between a pair of words using the

23
International Journal of Information and Education Technology, Vol. 12, No. 1, January 2022

Create Association Rules. The criterion used is confidence relatively higher rank compared to the other remaining
and the threshold value set for confidence is 0.90. campuses, tells us that most of the students on this campus
give greater importance to the usage of this term.
F. Flow Chart
TABLE IV: TOP 10 MOST OCCURRED TERMS FOR THE OVERALL DATA SET
COMPARED TO PER CAMPUS DATA SET

Alaminos Bayambang Binmaley Lingayen San Carlos Sta Maria Urdaneta


Word
rank freq. rank freq. rank freq. rank freq. rank freq. rank freq. rank freq.

105
good 1 547 1 344 1 421 1 940 1 841 1 391 1
2
100
teaching 2 389 2 215 2 363 2 932 2 679 2 295 2
7
students 3 312 4 101 3 154 3 504 3 323 3 192 3 783
topic 4 240 5 90 5 133 5 347 4 227 4 127 4 570
understand 11 131 10 47 8 104 4 391 5 200 5 110 8 440
class 6 146 8 52 4 135 7 337 6 182 10 83 11 369
time 21 76 11 45 17 67 19 196 7 168 21 47 7 462
teach 13 109 7 109 22 62 8 286 8 165 11 79 20 268
topics 12 111 17 32 24 52 12 253 16 122 13 57 10 388
teacher 8 138 3 122 9 98 16 215 10 161 7 98 35 150

B. Pattern Recognition Using FP-Growth


This research paper applied FP – Growth to determine the
frequent patterns in the data set. This operator efficiently
calculates all frequent itemsets from the given ExampleSet
using the FP-tree data structure. All attributes of the input
Fig. 3. The complete pattern recognition process. ExampleSet must be binominal.
FP – Growth has two basic working modes in identifying
Fig. 3 shows the process employed from data gathering up the most frequent itemset/s: 1) Looking for the least specified
until the use of RapidMiner to arrive at the desired results. number of itemsets with the highest support without
considering the 'minimum support', and 2) looking for every
itemset with support larger than the specified minimum
III. RESULTS AND DISCUSSION support. This approach uses the FP-Tree algorithm which
encodes the data set into a tree and then extracts the frequent
A. Descriptive Results itemsets from this tree. Frequent itemsets are groups of items
This subsection includes the terms with the highest that often appear together in the data.
frequency under the overall data set. Refer to Table III for the Using the RapidMiner 9.7, the dataset was generated and
top 10 most occurred terms from the overall data set. calculated based on FP - Growth method. The most frequent
items and itemsets were displayed in the Table V:
TABLE III: MOST OCCURRED TERMS FOR THE OVERALL DATA SET
Word Rank TABLE V: RESULTS AFTER USING FP-GROWTH FOR THE OVERALL DATA
good 1 SET
teaching 2 Size Support Item1 Item2 Item3 Item 4
students 3 3 0.031 good teaching she
topic 4 3 0.029 good teaching he
understand 5 3 0.012 good teaching students
class 6 3 0.009 good teaching topic
time 7 3 0.012 good teaching understand
teach 8 4 0.005 good teaching she understand
topics 9 4 0.004 good teaching he students
teacher 10 4 0.004 explain way things difficult
(min freq = 67, min items = 3, max items =5).
Based on the overall data, the term that occurred the most
is good. Identifying the rank of the top 10 terms provided The most frequent itemsets occurred are the following:
above concerning per campus provides a deeper itemset good and teaching for size 2 with support equal to
understanding of the data set. 0.141; itemset good, teaching, and she for size 3 with support
It can be gleaned from the table that the term good is still equal to 0.031; and itemset good, teaching, she, and
the top one terms occurred for all campuses. The San Carlos understand for size 4 with support equal to 0.005. For
City Campus almost reflected the actual rankings for the example, the terms good and teaching are the most common
overall data set except for the term topics. Table IV gives us terms used in the comments on the performance of the faculty
an idea of the terms mostly used by the students per campus members, that is, this item set occurs 1,410 times for every
when giving comments on the performance of the faculty 10,000 comments. Another, the itemset containing terms
members. Say, for example, the term teacher under the explain, way, things, and difficult means that it occurs 40
Bayambang Campus is ranked number three, which has a times for every 10,000 comments. The results in this section

24
International Journal of Information and Education Technology, Vol. 12, No. 1, January 2022

are useful in considering the rules that are worthy of further


TABLE XI: RESULTS AFTER USING FP-GROWTH FOR THE STA. MARIA
analysis. The FP -Growth results per campus are provided in DATA SET
the Table VI – XII. Also, Fig. 4 exhibits the comparison in Size Support Item1 Item2 Item3 Item 4
the support count of the most occurred terms among the 3 0.082 good he teaching
campuses of Pangasinan State University. 3 0.027 good he students
3 0.023 good he topic
TABLE VI: RESULTS AFTER USING FP-GROWTH FOR THE ALAMINOS DATA 3 0.075 good teaching she
SET 3 0.032 good teaching students
Size Support Item1 Item2 Item3 3 0.027 good teaching student
3 0.025 he good teaching 3 0.031 good teaching kind
3 0.005 he good students 3 0.027 good teaching attention
3 0.008 he good topic 3 0.023 good teaching abuse
3 0.021 good she teaching 3 0.023 good she students
3 0.007 good she students 3 0.023 good kind abuse
3 0.01 good she topic 3 0.025 good attention catch
3 0.006 topic can understand 3 0.023 he teaching kind
3 0.006 can explain topics 3 0.027 he topic understand
(min freq = 10, min items = 3, max items =5). 3 0.023 teaching kind abuse
4 0.023 good teaching kind abuse
TABLE VII: RESULTS AFTER USING FP-GROWTH FOR THE BAYAMBANG (min freq = 22, min items = 3, max items =5).
DATA SET
Size Support Item1 Item2 Item3 TABLE XII: RESULTS AFTER USING FP-GROWTH FOR THE URDANETA CITY
3 0.02 good teaching she DATA SET
3 0.017 good teaching he Size Support Item1 Item2 Item3 Item 4
3 0.01 good she teacher 3 0.021 he teaching good
3 0.011 good he instructor 3 0.009 he teaching students
3 0.014 communication skills social 3 0.006 he teaching topic
(min freq = 10, min items = 3, max items =5). 3 0.006 he teaching subject
3 0.018 teaching good she
TABLE VIII: RESULTS AFTER USING FP-GROWTH FOR THE BINMALEY 3 0.009 teaching good students
DATA SET 3 0.006 teaching good topic
Size Support Item1 Item2 Item3 Item 4 3 0.004 teaching good subject
3 0.046 good teaching she 4 0.003 he teaching good students
3 0.029 good teaching he 4 0.004 he topic can explain
4 0.009 good teaching she understand 4 0.004 teaching good she understand
4 0.01 good teaching dont like 4 0.004 she students topic understand
4 0.009 good topic giving examples 4 0.003 students topic subject understand
4 0.01 teaching she dont like (min freq = 15, min items = 3, max items =5).
4 0.009 she dont noisy like
(min freq = 10, min items = 3, max items =5).

TABLE IX: RESULTS AFTER USING FP-GROWTH FOR THE LINGAYEN DATA
SET
Size Support Item1 Item2 Item3 Item 4
3 0.033 she good teaching
3 0.007 she good students
3 0.012 she good understand
3 0.008 she good topic
4 0.008 she good teaching understand
4 0.009 she good teaching shes
4 0.006 she teaching understand shes
4 0.007 good teaching he students
4 0.006 good teaching voice low
(min freq = 20, min items = 3, max items =5).

TABLE X: RESULTS AFTER USING FP-GROWTH FOR THE SAN CARLOS CITY Fig. 4. Comparison of support count among the campuses of Pangasinan
DATA SET State University.

Size Support Item1 Item2 Item3 Item 4


3 0.022 good teaching she C. Pattern Recognition by Creating Association Rules
3 0.023 good teaching he Create Association Rules that are based on market basket
3 0.014 good teaching students analysis was used for pattern recognition. This operator
3 0.021 good teaching understand
3 0.015 good teaching class
generates a set of association rules from the given set of
3 0.016 she topic discuss frequent itemsets. This technique is often applied in the
3 0.014 explain way difficult supermarket and other shopping industries to identify the
3 0.015 explain way simple trending items that are purchased together; thus, contributing
3 0.014 explain difficult simple to the possible increase in profit of the said businesses by
3 0.014 way difficult simple
4 0.014 explain way difficult simple
determining the items that are subject to optimization or mass
(min freq = 32, min items = 3, max items =5). production. When this method is applied to text mining, each

25
International Journal of Information and Education Technology, Vol. 12, No. 1, January 2022

term is considered as an item and the text field is the resounding and clear (positive) or piercing (negative).
transactions. Hence, the goal is to further identify the terms
TABLE XIV: RESULTS AFTER CREATING ASSOCIATION RULES FOR PER
that occur together that will give us an overview of the CAMPUS DATA SET
perspective of the students towards their instructors. Campus Premises Conclusion Support Confidence
Applying the Create Association Rules operator gives a clear chance he 0.005867971 1.000000000
view and understanding of the terms that have a strong loud voice 0.006845966 1.000000000
relationship with each other and are commonly used together Alaminos strategies teaching 0.007334963 0.937500000
by the students in giving comments or evaluations to the lessons,
good 0.005867971 0.923076923
explaining
faculty members. The complete result for the overall data set social communication 0.013527575 1.000000000
is shown in the below: social skills 0.013527575 1.000000000
Bayambang
communication,
TABLE XIII: RESULTS AFTER CREATING ASSOCIATION RULES FOR THE skills 0.013527575 1.000000000
social
OVERALL DATA SET topic,
good 0.00969163 1.000000000
Premises Conclusion Support Confidence discussing
explain, way, discussing,
things 0.004309513 0.917808219 good 0.00969163 1.000000000
simple giving
explain, way, students,
difficult 0.004309513 0.917808219 understand 0.00969163 1.000000000
simple Binmaley situation
explain, way, noisy, like dont 0.014977974 1.000000000
simple 0.004309513 0.930555556
things like, messy dont 0.008810573 1.000000000
way, difficult explain, simple 0.004309513 0.943661972 she, noisy, like dont 0.008810573 1.000000000
explain, simple way 0.00469544 0.948051948 low voice 0.014096916 0.941176471
way, difficult explain, things 0.004373834 0.957746479 attend class 0.020264317 0.920000000
explain, things, good, teaching,
way 0.004373834 0.971428571 voice 0.0059988 1.000000000
difficult low
things, difficult explain 0.004502476 0.972222222 good, low voice 0.00989802 0.970588235
way, difficult, Lingayen
explain 0.004309513 1.000000000 soft voice 0.0074985 0.961538462
simple loud voice 0.016196761 0.947368421
(min freq = 67, min items = 1, max items =5, min confidence = 0.9). she, low voice 0.00869826 0.935483871
way, difficult explain 0.013864818 1.000000000
A total of 9 rules were derived wherein each numerical difficult, simple explain 0.014298094 1.000000000
value for confidence exceeds the set threshold value explain, way,
simple 0.013864818 1.000000000
(minimum confidence is 0.90) after applying the Create difficult
explain, simple way 0.014731369 0.971428571
Association Rules operator under the RapidMiner. This explain,
implies that the terms included in Table XIII have a strong simple 0.014298094 0.970588235
difficult
relationship with each other. For example, the terms way and San Carlos
explain,
way 0.013864818 0.969696970
difficult are strongly related with the terms explain and simple difficult, simple
simple way 0.01559792 0.947368421
with support equal to 0.004309513 and confidence equal to way, simple explain 0.014731369 0.944444444
0.943661972. Also, if way and difficult (premise) were explain, simple difficult 0.014298094 0.942857143
included in the comment, then there is a high possibility that explain, way,
difficult 0.013864818 0.941176471
the terms explain and simple (conclusion) will also appear. It simple
simple explain 0.015164645 0.921052632
is not appropriate to say that explain and simple (conclusion)
explain, simple way, difficult 0.013864818 0.914285714
occurred in the comment because of the existence of way and abuse good 0.023329799 1.000000000
difficult in the same comment. Take note that Association abuse teaching 0.023329799 1.000000000
Rules are not consequential; instead, they are co-existential, abuse kind 0.023329799 1.000000000
that is, when the premise occurred, then the conclusion will teaching,
good 0.026511135 1.000000000
attention
also occur at the same time. This is also because the result has abuse good, teaching 0.023329799 1.000000000
a low support value but with a high confidence value. Further, good, abuse teaching 0.023329799 1.000000000
this implies that the rules might not stand out in the data set teaching, abuse good 0.023329799 1.000000000
but they occurred frequently due to continued analysis. The abuse good, kind 0.023329799 1.000000000
same explanation goes for the rest of the rules included in the good, abuse kind 0.023329799 1.000000000
kind, abuse good 0.023329799 1.000000000
list. abuse teaching, kind 0.023329799 1.000000000
Conducting the same Association Rule process per campus
Sta Maria teaching, abuse kind 0.023329799 1.000000000
shown in Table XIV gives us the idea about the usual kind, abuse teaching 0.023329799 1.000000000
characteristics of its faculty members based on how the good, teaching,
abuse 0.023329799 1.000000000
kind
students perceived them. In Alaminos Campus, if lessons and
good, abuse teaching, kind 0.023329799 1.000000000
explaining (premise) were included in the comment, then teaching, abuse good, kind 0.023329799 1.000000000
there is a high possibility that the term good (conclusion) will good, teaching,
kind 0.023329799 1.000000000
also appear, hence most of the students perceived their abuse
instructors as good in explaining the lessons and critiqued kind, abuse good, teaching 0.023329799 1.000000000
good, kind,
them based on their teaching strategies. Also, if the term loud abuse
teaching 0.023329799 1.000000000
was in the comment, then there is a high possibility that the teaching, kind,
good 0.023329799 1.000000000
term voice will follow. This implies that students thought that abuse
their instructors are loud-voiced that may indicate whether teaching, good 0.026511135 0.961538462

26
International Journal of Information and Education Technology, Vol. 12, No. 1, January 2022

student this implies that the instructors have a profound


audible voice 0.004361371 1.000000000 understanding of the topics and lessons they are teaching,
inaudible voice 0.004361371 1.000000000
because their students deemed that they can thoroughly
faint voice 0.004153686 1.000000000
aids visual 0.007476636 1.000000000 explain things.
teaching, matter subject 0.003530633 1.000000000 In Sta Maria Campus, the students perceived their
depth, matter subject 0.003946002 1.000000000 instructors as good in teaching just like the other campuses
late, comes class 0.003115265 1.000000000 but abused them for being kind. This indicates that the
Urdaneta life, relates real 0.003322949 1.000000000
students recognized the kindness and potential of their
life, situations real 0.003738318 1.000000000
real, situations life 0.003738318 1.000000000 instructors in teaching, but their kindness made them
style teaching 0.003738318 0.947368421 vulnerable to abuse which imposes a serious concern. One of
he, reading materials 0.003738318 0.947368421 the things that an educator must possess is the ability to assert
soft, hear voice 0.003322949 0.941176471 authority inside the classroom. It is believed and through
explains, matter subject 0.005607477 0.931034483 experiences that too much authority inside the classroom
loud voice 0.017860852 0.914893617 affects the learning of the students and impedes innovation
(min items = 1, max items =5, min confidence = 0.9).
and feedback on the part of the students that is why
In Bayambang Campus, the majority of the students gave instructors need to set boundaries on when to show kindness
critic to their instructors through their communication and and when to impose strict authority. Thus, instructors in this
social skills because these terms have a strong relationship campus should exercise balance between imposing authority
with each other. Bayambang Campus is best known for its and showing kindness inside the classroom.
teacher education program because of having been granted Lastly, in Urdaneta Campus, most of the students
by The Accrediting Agency of Chartered Colleges and perceived their instructors as good in relating topics on
University in the Philippines (AACCUP) the Level IV status real-life situations to the subject matter. This might indicate
for Bachelor of Secondary Education and Bachelor of that the instructors are good at using an analogy in
Elementary Education until 2021; hence, good discussions; hence, this further indicates their thorough
communication and social skills are necessary for the understanding of the subject matter they are teaching. Also,
instructors to mold future educators and to maintain the students recognized their instructors whether loud-voiced or
excellence and quality teaching. soft-voiced akin to Lingayen Campus. Also, instructors are
As for Binmaley Campus, the students perceived their deemed by students as late in class or tardy. Tardiness in
instructors as good at discussing the topic. This implies that work implies irresponsibility and unreliability. Tardiness of
students are satisfied with the teaching methods employed, instructors breeds indifference, truancy or absenteeism in the
and the knowledge and pieces of information delivered by students. Instructors must take their attendance and
their instructors. Also, students perceived their instructors as punctuality seriously so as to foster a good classroom
understanding especially when it comes to the situation of environment and to be role models to the students.
their students. The only downfall is that students perceived It is meaningful and just to introduce another parameter
their instructors as having low voice quality. Hence, it since the support value is low with a high confidence value to
explains why they do not like noise particularly for female measure the effectiveness of the association rules - which is
instructors due to the relationship between the terms she, the lift. The lift is the ratio between the support and the
noisy, like to don’t. product of the probabilities of the premise and the conclusion
In Lingayen Campus, if the terms good, teaching, and low (or the left and right-hand side) occurring assuming that there
appeared in the comment, then the term voice will appear at is no association between them or independent from each
the same time. This implies that the majority of students other. The lift measures the strength of the association
recognized their instructors as good in teaching but with a between the premise and conclusion; hence, the larger the
low voice, especially when the instructor is female because of value of the lift the stronger the association between them. A
the strong relationship between the terms she, low, and voice. lift value equal to 1 suggests that there is no meaningful
Also, the terms loud and voice that appeared on the result association between the premise and conclusion. Notice that
asserts the existence of instructors that possess audible voices. every lift value of each of the following derived rules is
Just like in Alaminos Campus, this may indicate whether greater than 1; thus, there is a strong association between the
resounding and clear (positive) or piercing (negative). The premise and conclusion. Table XV exhibits the top three rules
ambiguity of the usage of loud and voices is probably due to with the strongest association per campus highlighted in
the insufficiency of objects and complements in the yellow.
comments of the students, thus, the majority of the comments TABLE XV: LIFT VALUE OF RULES PER CAMPUS
are deemed as incomplete sentences. Campus Premises Conclusion Lift
In San Carlos Campus, most of the students perceived that chance he 3.205329154
their instructors can explain difficult topics, lessons, and loud voice 49.87804878
other similar things thoroughly since the terms simple and Alaminos strategies teaching 4.992675781
lessons,
way have a strong relationship, also the itemset simple, way, explaining
good 3.609354317
difficult, and explain has a strong relationship. This is a good social communication 41.7826087
skill for any educator in facilitating the understanding and social skills 41.7826087
Bayambang
acquisition of knowledge among the students. Furthermore, communication,
skills 41.7826087
social

27
International Journal of Information and Education Technology, Vol. 12, No. 1, January 2022

topic, discussing good 2.844611529 members. Further, this study aims to determine the most
discussing, giving good 2.844611529 frequent word/words to describe the performance of a faculty
students, situation understand 11.58163265
member and to determine the performance of faculty
noisy, like dont 12.20430108
Binmaley members of every campus of Pangasinan State University
like, messy dont 12.20430108
she, noisy, like dont 12.20430108 using the associative approach text analytics technique.
low voice 22.72841051 RapidMiner9.7. was used to evaluate the results in this study.
attend class 8.420967742 All members of this paper evaluated the results manually to
good, teaching,
low
voice 12.97276265 ensure validity.
good, low voice 12.5912108 The result reveals that the term good (excluding the
Lingayen
soft voice 12.47381024 pronouns he and she) is the most occurring term on the
loud voice 12.28998566 students' evaluation using associative rule analysis. It can be
she, low voice 12.13581022 deduced that the majority of the teaching staff of Pangasinan
way, difficult explain 19.23333333
difficult, simple explain 19.23333333
State University across every campus are good at teaching
explain, way, their subject or topic and can explain it simply so that
simple 60.73684211
difficult students can easily comprehend it.
explain, simple way 22.42057143 It is revealed that "teaching good subject/topic," "explain
explain, difficult simple 58.9504644
simple," and other meaningful associated words can be
explain, difficult,
way 22.38060606 utilized to evaluate the performance of the teacher. The
simple
simple way 21.86526316 results reveal not only the quantitative values of faculty
way, simple explain 18.16481481 evaluation it also exposed the qualitative opinion of the
explain, simple difficult 57.26616541 students in the performance of their faculty members.
explain, way,
simple
difficult 57.16408669 This study reveals important aspects of the teaching
simple explain 17.71491228 performance of the faculty members of Pangasinan State
explain, simple way, difficult 65.94285714 University. The results can be used for coaching and
teaching, attention good 17.46296296 mentoring by university and campus heads to their faculty
abuse good, teaching 2.619444444
members in terms of their weaknesses. Moreover, the results
good, abuse teaching 4.668316832
San Carlos
teaching, abuse good 3.367857143 can be utilized by Pangasinan State University to evaluate the
abuse good, kind 2.619444444 teaching performance of their faculty members based on the
good, abuse kind 27.73529412 comments or opinions of the students.
kind, abuse good 17.46296296
abuse teaching, kind 2.619444444
CONFLICT OF INTEREST
teaching, abuse kind 29.46875
kind, abuse teaching 17.46296296 The authors declare no conflict of interest.
good, teaching,
abuse 3.367857143
kind
AUTHOR CONTRIBUTIONS
good, abuse teaching, kind 32.51724138
teaching, abuse good, kind 29.46875 F.F.P., P.V.C., B.F.R., J.M.P. conceived the presented idea
good, teaching, of this paper; P.V.C. and J.M.P. gathered every necessary
kind 27.73529412
abuse
data; D.B.A.G. conducted the preprocessing of data; F.F.P.
kind, abuse good, teaching 17.46296296
good, kind, abuse teaching 4.668316832 and P.V.C. performed the analysis; F.F.P., P.V.C., B.F.R.,
teaching, kind, and J.M.P. supervised the findings of this work. All authors
good 3.367857143
abuse stated on this paper discussed the results and contributed to
teaching, student good 2.619444444 the completion of the final manuscript.
audible voice 12.57180157
inaudible voice 12.57180157
faint voice 12.57180157 REFERENCES
aids visual 102.4468085 [1] J. P. Sampson, M. P. Driscoll, D. F. Foulk, and P. S. Carroll,
teaching, matter subject 11.12009238 “Successful faculty performance in teaching, research and original
depth, matter subject 11.12009238 creative work, and service,” 2010.
late, comes class 13.67897727 [2] N. Temmerman, “Transforming higher education in Vietnam,”
Urdaneta life, relates real 63.35526316 University World News, 2019.
life, situations real 63.35526316 [3] G.W. Fry, “Higher education in Vietnam,” The Political Economy of
Educational Reforms and Capacity Development in Southeast Asia,
real, situations life 54.1011236
Education in the Asia-Pacific Region: Issues, Concerns, and Prospects,
style teaching 4.612314406
vol 13. Springer, Dordrecht. 2009.
he, reading materials 52.43194192 [4] A. K. Pal and S. Pal, “Evaluation of teacher's performance: A data
soft, hear voice 11.83228383 mining approach,” International Journal of Computer Science and
explains, matter subject 10.35318946 Mobile Computing, vol. 2, issue 12, pp. 359-369, 2013.
loud voice 11.50186101 [5] C. Magno, A Meta-Evaluation Study on the Assessment of Teacher
Performance in an Assessment Center in the Philippines, vol. 3, 2003.
[6] R. I. Miller, Evaluating Faculty for Promotion and Tenure. The Jossey
Bass Higher Education Series, Jossey-Bass Publishers, 350 Sansome
IV. CONCLUSION Street, San Francisco, CA 94104, 1987.
[7] W. L. Sanders and S. P. Horn, “Research findings from the Tennessee
This research aims to develop a model to predict the Value-Added Assessment System (TVAAS) database: Implications for
performance of the faculty using associative rule based on the educational evaluation and research,” Journal of Personnel Evaluation
existing evaluation form use by PSU to evaluate faculty in Education, vol. 12, pp. 247-256, 1998.

28
International Journal of Information and Education Technology, Vol. 12, No. 1, January 2022

[8] A. Mustafa, Predicting Instructor Performance Using Data Mining Paulo V. Cenas received his M.A. applied statistics
Techniques in Higher Education, 2016. degree from Benguet State University and Ed.D.
[9] S. Badur and B. Badur, “Analyzing teaching performance of instructors Mathematics degree from Pangasinan State University.
using data mining techniques,” 2011. Currently, he is the vice president for research,
[10] R. S. J. D. Baker and K. Yacef, “The state of educational data mining in extension, and innovation at Pangasinan State
2009: A review and future vision,” J. Edu. Data Mining, vol. 1, no. 1, University. His current research interests include
pp. 3-17, 2009. system development, digital learning, and business
[11] F. D. Kentli and Y. Sahin, “An SVM approach to predict student management.
performance in manufacturing processes course,” Energy, Edu., Sci.
Technol. B, vol. 3, no. 4, pp. 535-544, 2011. Bobby F. Roaring received his M.A. applied statistics
[12] A. Pena-Ayala, “Review: Educational data mining: A survey and a data degree from Benguet State University. He serves as the
mining-based analysis of recent works,” Expert Syst. Appl., vol. 41, no. head of the Statistics Center at Pangasinan State
4, pp. 1432_1462, 2014. University. Mr. Roaring's main research interests
[13] J. Zimmerman, K. H. Brodersen, H. R. Heinimann, and J. M. include machine learning and data mining.
But-unarm, “A model-based approach to predicting graduate-level
performance using indicators of undergraduate-level performance,” J.
Edu. Data Mining, vol. 7, no. 3, pp. 151-176, 2015.

Copyright © 2022 by the authors. This is an open access article distributed Jennifer M. Parrone received her M.S. in mathematics
under the Creative Commons Attribution License which permits unrestricted at the University of the Philippines (UP) Baguio City.
use, distribution, and reproduction in any medium, provided the original She serves as the chair of the Mathematics and Natural
work is properly cited (CC BY 4.0). Sciences Department at Pangasinan State University –
Urdaneta City Campus. Her main research interests
Frederick F. Patacsil received his master of computer include digital learning and game-based learning.
science and doctor in information technology degree
from the University of the Cordilleras in 2003 and 2016
respectively. Since 1994, he has been with Pangasinan Daniel Bezalel A. Garcia is an M.S. Mathematics
State University, where he is currently a faculty at the student at the University of the Philippines (UP)
Department of Information Technology. His current Baguio City. He is a faculty member of the
research interests include data mining, software Mathematics and Natural Sciences Department at
development, and machine learning. Pangasinan State University – Urdaneta City Campus.
Mr. Garcia’s main research interests include parameter
estimation, digital learning, and data mining.

29

You might also like