Appraisal Evaluation Using Fuzzy Logic
Phambahadur Pun
Department of Computer Science
University of Huddersfield
Huddersfield, United Kingdom
u1976481@unimail.hud.ac.uk
Abstract—This electronic document is a “live” template and
already defines the components of your paper [title, text, heads,
etc.] in its style sheet. *CRITICAL: Do Not Use Symbols, Special
Characters, Footnotes, or Math in Paper Title or Abstract.
(Abstract)
Keywords—component, formatting, style, styling, insert (key
words)
I. PROBLEM STATEMENT
A Dorset-based organisation called Collaborative
Design Team (CDT) has been exploring options to
evaluate its employee’s performance using disruptive
technology of Artificial Intelligence. The company
provides Annual Appraisal Reports (AAR) to evaluate
its employees’ performance which is used for
promotion, posting and performance enhanced bonus
scheme. In the last few years, few disgruntled
employees have raised complaints doubting the
integrity and authenticity of the performance evaluation
process. Upon investigating the complaints, the
following key facts were identified:

Appraisal reports were written subjectively
rather than objectively and contained distinctly
identifiable personal and sensitive information.

The performance evaluation board members
were pooled from the same organisation and
there was a possibility of an unconscious bias.

The grades awarded on AARs were random
and often didn’t marry up with the written
evidence on performance and potential
paragraphs.
II. RESEARCH
A. Structure of AAR
A solution is urgently required to rectify the concerns
raised by CDT’s unhappy employees. The proposed
solution must be trialled and evaluated and certified by
the company administrator before implementing it as a
solution. The AAR format has been broken down into
five sections as shown in fig 1.
Figure 1: Annual Appraisal Report Sample
1) Employee
Detail. This section contains the
employee’s name, Post Identification Number (PID),
current grade and total service years completed. It is
recommended that the service number or randomly
generated number is used for the promotion board for
anonymity.
2) Attributes. This is the most important section of the
AAR which has six attributes in total. These attributes
have three grades: below average, average and high
Qualities
Initiatives
Grades
Below
average,
average
or high
Scores
1-10
Communication
Power
Results
Same as
above
Same as
above
Same as
above
Same as
above
Same as
above
1-10
Leadership
Problem
Solving
Decision
Making
1-10
1-10
1-10
1-10
Remarks
Below
average = 14
Average =
4-8
High = 8-10
Same
as
above.
Same
as
above.
Same
as
above.
Same
as
above.
Same
as
above.
Table 1: Attributes and their Grading Criteria
with numerical grades between 0 and 10 assigned next
to it. These grades will be used as the primary source
of information to evaluate the employees’ performance.
Table 1 shows the classification of grades for each
attribute.
3) Performance. This is a free text section that allows
up to 120 words to describe the performance of an
employee. This should act as an evidence of the grades
awarded to six attributes. For this assignment, this
section won’t be used for evaluation as it requires a
XXX-X-XXXX-XXXX-X/XX/$XX.00 ©20XX IEEE
sophisticated Natural Language Processing (NLP)
algorithm.
4) Potential. This is a free text section that allows up to
120 words to describe the future potential of an
employee. This section is more useful for the allocation
of posting once the candidate is successful on the
promotion board. For this assignment, this section won’t
be used for evaluation as it requires a sophisticated
NLP algorithm.
5) Recommendations. This section allows the
reporting officer to categorically state the candidate’s
promotability and future employment. Future
employment can be either management, human
resources or technical.
III. METHODOLOGY
Figure 3: Mamdani Output Membership Function (MATLAB
2021)
2. Fuzzy operator. Both Mamdani and Sugeno FIS were
used. The Mamdani FIS is the most widely used and it is
easier to set up. The output membership function of Sugeno
is represented by a constant parameter as shown in fig 4.
FIS and ANFIS techniques will be used to predict the
overall score using six attributes as the inputs. It is
recommended that the service number or randomly
generated numbers be used instead of name to
maintain anonymity.
A. FIS using MATLAB
MATLAB’s Fuzzy Logic Designer Toolbox (MATLAB, 2021)
is used to create membership function and rules. Fig 2 shows
that it is using Mamdani and has six inputs and one output.
Figure 4: Sugeno Output Membership Function (MATLAB
2021)
3. Rules. Since there are six inputs with three membership
functions (MFs), it would require (3x3x3x3x3x3) 729 rules in
total. A specific rules can be defined here. For example, the
company doesn’t want to promote people if their performance
is below average in at least one area. The first row of table 2
has an output of 1 because the leadership quality is below
average despite qualifying on the other five attributes.
Figure 2:Fuzzy Logic Designer (MATLAB 2021)
1. Inputs and Outputs. Each input has a range of 0-10 where
0-4 represents below average, 4-8 represents the average
and 8-10 represents high. The output has a range of 0-10
where 0-4, 4-8 and 8-10 represent ‘can’t promote’, ‘could
promote’ and ‘should promote’ respectively as shown in fig 3.
L1
L2
L3
L4
L5
L6
Output
3
2
2
1
3
1
1
3
2
2
2
3
2
2
3
2
3
2
2
3
2
3
3
3
3
3
3
3
3
1
1
2
1
1
Table 2: Example of FIS Rules
1
4. Evaluation. The pattern of Mamdani and Sugeno FIS
outputs were found to be almost identical as demonstrated by
figure 5. Whilst testing with few sample inputs, both FIS
methods produced an accurate output due to 729 rules
defined earlier.
Table 4: ANFIS Error Comparison
Mamdani Vs Sugeno Output
Graph
10
0
2. Evaluation. The plot of sub clustering output against the
checking dataset is shown in fig 6. The epochs figure was
gradually increased to see if it reduced the errors but it didn’t
have any impact.
Mamdadni Output
123 4 5 6
7 8 9 10
11 12
Mamdadni Output
Suzeno Output
Figure 5: Comparison of Mamdani and Sugeno Outputs
ANFIS using MATLAB
Mamdani and Sugeno FIS methods are simple and effective
but they become complicated when input variable and
membership function increase significantly. Adaptive NeuroFuzzy Inference System (ANFIS) is a class of adaptive
networks that incorporate both neural networks and fuzzy
logic principles. A neural network is a supervised learning
algorithm that utilises a historical dataset to predict the future
outcome (Mathur, Glesk, & Buis, 2016). For this assignment,
MATLAB’s Neuro-Fuzzy Designer application is used.
1. Training. The historical training dataset is the most
important element of an ANFIS model because the predicted
outcome relies on it. The company provided a historical
dataset from last year’s promotion board. After analysing the
data, it was evident that the company only penalised
candidates who scored below average in Problem Solving and
Decision Making category. The candidates were given three
grades as shown in table 3.
Attributes
Initiatives
Scores
1-3
Communication
1-3
Power
Results
1-3
Leadership
1-3
Problem Solving
1-3
Decision Making
1-3
Table 3: Input and MFs for ANFIS
Remarks
Below average = 1
Average = 2
High = 3
Same as above.
Same as above.
Same as above.
Same as above.
Same as above.
The training, testing and checking datasets are attached to
annex A, were used to train the ANFIS model using various
configurations as shown in table 4. Sub clustering FIS type
produced the least number of errors.
Input FIS Type
Output
Testing
Testing
(No
MF Type
Error
Error
of
with
with
MFs)
testing
checking
data
data
3 3 3 Grid
trim
1.7759
1.4753
333
partition
Constant
trim inear
1.7759
1.4753
3 3 3 Grid
gbellmf
1.409
1.1471
333
partition
constant
gbellmf
1.4326
1.1933
linear
3 3 3 Grid
constant
1.4326
1.1702
333
partition
linear
1.4442
1.18
0.4254
0.4369
3 3 3 Sub
N/A
333
clustering
Figure 6: ANFIS Sub Clustering Plot (MATLAB 2021)
3. Findings. Although the error rate remains high, the sub
clustering ANFIS model was selected as the solution.
Strangely, It produced 49 membership functions for each input
instead of three. From fig 6, it can be seen that the accuracy
slips as the output increases towards 3. The training dataset
had only 49 records. Ideally, it needs to be bigger than that
and it must include data from multiple years. Both online and
offline version of MATLAB software was found to be very slow
and often crashed during the testing phase. To alleviate this
problem, a script and dashboard were created which
significantly improved the testing experience.
4. MATLAB Dashboard. Script_Mam.m allows data input
through the MATLAB command window without running the
.fis file. The script was further improved by creating a
dashboard using features from Open-Source MATLAB GUI (
SysMat Soft Solutions, 2021). Fig 7 shows a screenshot of the
dashboard which can accept manual data, bulk data, export
results and plot a 3D surface area. It can support Mumdani,
Sugeno and ANFIS model as long as the .fis file is changed
accordingly on the Main.m file.
Times New Roman or the Symbol font (please no other font).
To create multileveled equations, it may be necessary to treat
the equation as a graphic and insert it into the text after your
paper is styled.
Number equations consecutively. Equation numbers,
within parentheses, are to position flush right, as in (1), using
a right tab stop. To make your equations more compact, you
may use the solidus ( / ), the exp function, or appropriate
exponents. Italicize Roman symbols for quantities and
variables, but not Greek symbols. Use a long dash rather than
a hyphen for a minus sign. Punctuate equations with commas
or periods when they are part of a sentence, as in:

Figure 7: MATLAB Dashboard (MATLAB 2021)
Conclusions and Recommendations
Mamdani FIS is the easiest system to design and implement
but requires a complex set of rules. When the inputs and
membership functions are too big and complex, ANFIS is
undoubtedly the better choice. The main advantage of ANFIS
is that it can learn from the historical training dataset and can
adapt constantly to predict an output.
AAR is a complex piece of document which contains attributes
and a comprehensively written narrative on the performance
and potential of an individual. This assignment has not delved
into the evaluation of written narratives which can be analysed
by NLP techniques such as Random Forest.
Although the proposed ANFIS model provides a workable
solution, it needs further investigation before it can be
implemented by the company. It can however be used as a
tool alongside the human performance evaluation board.
IV. REFERENCES
SysMat Soft Solutions. (2021, April 24). Retrieved from Free
Thesis: https://free-thesis.com/product/matlab-guipresentation-for-fuzzy-logic-application/
ab

Note that the equation is centered using a center tab stop.
Be sure that the symbols in your equation have been defined
before or immediately following the equation. Use “(1)”, not
“Eq. (1)” or “equation (1)”, except at the beginning of a
sentence: “Equation (1) is . . .”
B. Some Common Mistakes
 The word “data” is plural, not singular.
 The subscript for the permeability of vacuum 0, and
other common scientific constants, is zero with
subscript formatting, not a lowercase letter “o”.
 In American English, commas, semicolons, periods,
question and exclamation marks are located within
quotation marks only when a complete thought or
name is cited, such as a title or full quotation. When
quotation marks are used, instead of a bold or italic
typeface, to highlight a word or phrase, punctuation
should appear outside of the quotation marks. A
parenthetical phrase or statement at the end of a
sentence is punctuated outside of the closing
parenthesis (like this). (A parenthetical sentence is
punctuated within the parentheses.)
 A graph within a graph is an “inset”, not an “insert”.
The word alternatively is preferred to the word
“alternately” (unless you really mean something that
alternates).
 Do not use the word “essentially” to mean
“approximately” or “effectively”.
Mathur, N., Glesk, I., & Buis, A. (2016). Comparison of
adaptive neuro-fuzzy inference system (ANFIS).
Glasgow: University of Strathclyde.
 In your paper title, if the words “that uses” can
accurately replace the word “using”, capitalize the “u”;
if not, keep using lower-cased.
Mathworks. (2021, February 04). MATLAB R2020b Software.
 Be aware of the different meanings of the homophones
“affect”
and
“effect”,
“complement”
and
“compliment”, “discreet” and “discrete”, “principal”
and “principle”.
MATLAB. (2021, April 25). Adaptive Neuro-Fuzzy Inference
System.
Retrieved
from
https://uk.mathworks.com/help/fuzzy/train-adaptiveneuro-fuzzy-inference-systems-gui.html
 Do not confuse “imply” and “infer”.
 The prefix “non” is not a word; it should be joined to
the word it modifies, usually without a hyphen.
A. Equations
The equations are an exception to the prescribed
specifications of this template. You will need to determine
whether or not your equation should be typed using either the
 There is no period after the “et” in the Latin
abbreviation “et al.”.
 The abbreviation “i.e.” means “that is”, and the
abbreviation “e.g.” means “for example”.
An excellent style manual for science writers is [7].
V. USING THE TEMPLATE
After the text edit has been completed, the paper is ready
for the template. Duplicate the template file by using the Save
As command, and use the naming convention prescribed by
your conference for the name of your paper. In this newly
created file, highlight all of the contents and import your
prepared text file. You are now ready to style your paper; use
the scroll down window on the left of the MS Word
Formatting toolbar.
A. Authors and Affiliations
The template is designed for, but not limited to, six
authors. A minimum of one author is required for all
conference articles. Author names should be listed starting
from left to right and then moving down to the next line. This
is the author sequence that will be used in future citations and
by indexing services. Names should not be listed in columns
nor group by affiliation. Please keep your affiliations as
succinct as possible (for example, do not differentiate among
departments of the same organization).
1) For papers with more than six authors: Add author
names horizontally, moving to a third row if needed for more
than 8 authors.
2) For papers with less than six authors: To change the
default, adjust the template as follows.
a) Selection: Highlight all author and affiliation lines.
b) Change number of columns: Select the Columns
icon from the MS Word Standard toolbar and then select the
correct number of columns from the selection palette.
c) Deletion: Delete the author and affiliation lines for
the extra authors.
B. Identify the Headings
Headings, or heads, are organizational devices that guide
the reader through your paper. There are two types:
component heads and text heads.
Component heads identify the different components of
your paper and are not topically subordinate to each other.
Examples include Acknowledgments and References and, for
these, the correct style to use is “Heading 5”. Use “figure
caption” for your Figure captions, and “table head” for your
table title. Run-in heads, such as “Abstract”, will require you
to apply a style (in this case, italic) in addition to the style
provided by the drop down menu to differentiate the head
from the text.
Text heads organize the topics on a relational, hierarchical
basis. For example, the paper title is the primary text head
because all subsequent material relates and elaborates on this
one topic. If there are two or more sub-topics, the next level
head (uppercase Roman numerals) should be used and,
conversely, if there are not at least two sub-topics, then no
subheads should be introduced. Styles named “Heading 1”,
“Heading 2”, “Heading 3”, and “Heading 4” are prescribed.
C. Figures and Tables
a) Positioning Figures and Tables: Place figures and
tables at the top and bottom of columns. Avoid placing them
in the middle of columns. Large figures and tables may span
across both columns. Figure captions should be below the
figures; table heads should appear above the tables. Insert
figures and tables after they are cited in the text. Use the
abbreviation “Fig. 1”, even at the beginning of a sentence.
TABLE I.
Table
Head
copy
TABLE TYPE STYLES
Table Column Head
Table column subhead
Subhead
Subhead
More table copya
a.
Sample of a Table footnote. (Table footnote)
Fig. 1. Example of a figure caption. (figure caption)
Figure Labels: Use 8 point Times New Roman for Figure
labels. Use words rather than symbols or abbreviations when
writing Figure axis labels to avoid confusing the reader. As an
example, write the quantity “Magnetization”, or
“Magnetization, M”, not just “M”. If including units in the
label, present them within parentheses. Do not label axes only
with units. In the example, write “Magnetization (A/m)” or
“Magnetization {A[m(1)]}”, not just “A/m”. Do not label axes
with a ratio of quantities and units. For example, write
“Temperature (K)”, not “Temperature/K”.
ACKNOWLEDGMENT (Heading 5)
The preferred spelling of the word “acknowledgment” in
America is without an “e” after the “g”. Avoid the stilted
expression “one of us (R. B. G.) thanks ...”. Instead, try “R.
B. G. thanks...”. Put sponsor acknowledgments in the
unnumbered footnote on the first page.
REFERENCES
The template will number citations consecutively within
brackets [1]. The sentence punctuation follows the bracket [2].
Refer simply to the reference number, as in [3]—do not use
“Ref. [3]” or “reference [3]” except at the beginning of a
sentence: “Reference [3] was the first ...”
Number footnotes separately in superscripts. Place the
actual footnote at the bottom of the column in which it was
cited. Do not put footnotes in the abstract or reference list. Use
letters for table footnotes.
Unless there are six authors or more give all authors’
names; do not use “et al.”. Papers that have not been
published, even if they have been submitted for publication,
should be cited as “unpublished” [4]. Papers that have been
accepted for publication should be cited as “in press” [5].
Capitalize only the first word in a paper title, except for proper
nouns and element symbols.
For papers published in translation journals, please give
the English citation first, followed by the original foreignlanguage citation [6].
[1]
[2]
[3]
[4]
[5]
G. Eason, B. Noble, and I. N. Sneddon, “On certain integrals of
Lipschitz-Hankel type involving products of Bessel functions,” Phil.
Trans. Roy. Soc. London, vol. A247, pp. 529–551, April 1955.
(references)
J. Clerk Maxwell, A Treatise on Electricity and Magnetism, 3rd ed.,
vol. 2. Oxford: Clarendon, 1892, pp.68–73.
I. S. Jacobs and C. P. Bean, “Fine particles, thin films and exchange
anisotropy,” in Magnetism, vol. III, G. T. Rado and H. Suhl, Eds. New
York: Academic, 1963, pp. 271–350.
K. Elissa, “Title of paper if known,” unpublished.
R. Nicole, “Title of paper with only first word capitalized,” J. Name
Stand. Abbrev., in press.
[6]
[7]
Y. Yorozu, M. Hirano, K. Oka, and Y. Tagawa, “Electron spectroscopy
studies on magneto-optical media and plastic substrate interface,” IEEE
Transl. J. Magn. Japan, vol. 2, pp. 740–741, August 1987 [Digests 9th
Annual Conf. Magnetics Japan, p. 301, 1982].
M. Young, The Technical Writer’s Handbook. Mill Valley, CA:
University Science, 1989.
We suggest that you use a text box to insert a graphic
(which is ideally a 300 dpi TIFF or EPS file, with all fonts
embedded) because, in an MSW document, this method is
somewhat more stable than directly inserting a picture.
To have non-visible rules on your frame, use the
MSWord “Format” pull-down menu, select Text Box >
Colors and Lines to choose No Fill and No Line.
IEEE conference templates contain guidance text for
composing and formatting conference papers. Please
ensure that all template text is removed from your
conference paper prior to submission to the
conference. Failure to remove template text from
your paper may result in your paper not being
published.