This document outlines a proposed system for evaluating employee performance. It describes collecting data on various indicators like attendance, customer feedback, and task completion. It then discusses using a decision tree algorithm to generate rules for evaluating performance based on the data. A fuzzy logic system is also proposed to map input data to linguistic variables and output an evaluation. The goal is to develop a more objective, data-driven approach to performance reviews.
3. Introduction
• Evaluate employee on basis of the various
indicators.
• With the development of this system we will
find an interesting and better way for knowing
about the performances.
• The various indicators are attendance,
customer feedback, task completion, dealing
capacity, etc. Once all the data's are fetched in
the system, it rates the employees.
4. Problem with existing system
• Manual
• Not Potential candidate
• Focusing on Negative
• Rating in Middle
• Ruins teamwork and team spirit.
5. Objective
• To evaluate the performance of any employee based
on indicators.
• A performance evaluation is a constructive process
to acknowledge an employee’s performance.
• to improve the employee's contributions to the
organization's goals.
• This will help in increasing and improving the
efficiency and productivity of any organization
9. Methodology
Data collection
• We collected data from managers , sub
managers ,supervisor e.t.c
• We gone through all the reports
• We consult with 360 feedback receiver
• We consult with all the co wokers , staffs as
well as consumers
10. Types of data collection
• Third party feedback
–360 survey
• Competencies
–Sincerity
11. Repot submission
Attendance
Creativity
Reliability
Total client handle
Total profit made
Sincerity
360 degree survey
Goal progress generic data
Competencies third party
feedback
Data classification
12. Algorithm :-ID3
• If every element in the subset belongs to the
same class , then the node is turned into a leaf
and labeled with the class of the examples
• If the examples do not belong to the same class
• Calculate entropy and hence information gain to
select the best node to split data.
• Partition the data into subset.
• Repeat
13. Reason for Decision Tree
• Decision trees implicitly perform variable
screening or feature selection
• Decision trees require relatively little effort
from users for data preparation
• Nonlinear relationships between parameters do
not affect tree performance
• The best feature of using trees for analytics -
easy to interpret and explain to executives!
14.
15.
16.
17. Rule generation
• Rule base classifier
• using decision tree for generating rule
• Extract rule from decision tree
20. Fuzzy Logic
• fuzzy logic
• Fuzzifier : used to convert crisp vale into
linguistic form
• Knowledge base : where knowledge is stored
• Inference engine : used to map the rules ,
priority base.
• Defuzzifier : used to convert linguistic variable
into crisp value
22. INPUT INPUTNAME LINGUSTIC RANGE
Input1
Attendance
Bad 1-50
Good 25-75
VG 50-100
Input2 Total client handle Bad 1-50
Good 25-75
VG 50-100
Input3 Total profit made
Bad 1-501-50
Good 25-75
VG 50-100
Input4 Sincerity
Bad 1-50
Good 25-75
Very good 50-100
Input5 Feedback
Bad 1-50
Good 25-75
Very good 50-100
Input6 Report submission
Bad 1-50
Good 25-75
Very good 50-100
Input7 Reliability
Bad 1-50
Good 25-75
Very good 50-100
Input8 Creativity
Bad 1-50
Good 25-75
Very good 50-100