CO528 Introduction to Intelligent Systems: Genetic Algorithm Solution

This document presents a comprehensive solution to a genetic algorithm assignment for the CO528 Introduction to Intelligent Systems course at the University of Kent. The assignment focuses on predicting results from a dataset using a genetic algorithm implemented in Java. The solution includes the Java source code, a detailed description of the algorithm with a flowchart, and a justification for choosing the genetic algorithm approach. The student also provides an estimation of the prediction accuracy, including the methodology for splitting the labeled data into training and testing sets. The document also contains a table with predicted values for the second dataset and a discussion on the impact of dataset size and splitting on the algorithm's performance. The assignment demonstrates the application of genetic algorithms in addressing prediction tasks, offering insights into the algorithm's implementation, performance evaluation, and practical considerations in data splitting.

GENETIC ALGORITHAM
[Type the document subtitle]
[Pick the date]
[Type the company name]
i

Table of Contents
1)................................................................................................................................................3
2)................................................................................................................................................3
3)................................................................................................................................................3
B.1..........................................................................................................................................4
B.2..........................................................................................................................................4
B.3..........................................................................................................................................5
2 | P a g e

1) Here ,I have choose genetic algorithm
2) Here, I have selected the java programming language. It is an object-oriented programming
language and easy to understand and run. I have use class and method for developed genetic
algorithm and easy to implement machine programming and prediction algorithm.
Java helps to write code with methods and functions. It provides the reusability of the code which
makes the program more flexible and reliable.
3)
3 | P a g e

B.
Genetic Algorithms are algorithms that are designed for large areas of search space. It helps to find
the best solution by natural selection and evolution. A summary of possible solutions to find better
solutions to value and get better conclusions.
Our solution solve
Number of permanent models.
Edit processing time
We've found a good solution / solution.
Kill your manual work.
B.1
Flow chart
B.2
we choose Genetic algorithms because it is more effective while we have little information about the
field of research. If you know what solution you need. But you have no idea how to get a problem
solution. GA is the best solution. Here we focus on how to fix recurring growth / common problems
and provide solutions for competitiveness, capture changes, and more genetics than humans. In
solving GA problems, you must choose the appropriate way to a method to find a solution to the
problem you are facing. Therefore, you can create a randomized solution to create a beginner size
group and fit your fitness. the specific solution to a problem or increase the population by changing
the cross factors, etc. and choose a solution to build the next generation community using the
4 | P a g e

method. It should be repeated. Exceeding the correct number of repetitions. This solves the problem
with a good problem-solving set of results.
B.3
Here, we have checked the data is large and we need to implement the genetic algorithm. Here data
set is available in a text file and work on the dataset we used the prediction algorithm. We have split
the dataset 20-80 percentage among the training and testing stage. Here, we figure out how much
data we have split and make the validation of the dataset. If the training set is small, in this case, the
algorithm will not enough to learning effectively. The set of validation is small and there is variance
in the accuracy and precision will be large. Hence we split the label to the training set and validation
good place to start. The data set is small, splitting to 80/20 points, which can cause a lot of
variances. we write the code for the evaluate the N fold which is called as cross-over (Checking). Our
main purpose is we will complete all the steps with time and accuracy. For example, when validating
10 times, we will check up to ten percent of the information, find statistics. Again computing these
statistics, repeat ten times. We check the dataset chunk. training has 99% accuracy in training and
testing. Expected to be lower than the test set (Before sharing data, we do not care about the same
spam repetition). We have trained test data, so we don't need to measure the format of new
measurement data anymore.
5 | P a g e

We split the data set into training sets. The information contained in the package of training
means labels. This set is used to create our model. Models learn from data sets with these tags and
give a general understanding of the relationship of data in a data set. We usually divide the data set
according to the 80/20 rule. That is 80% of the data set enters the package of training where 20% of
the test set.
B.4
Result for 3533450_203248480_assignement5281
6 | P a g e

7 | P a g e