National Action Council for Minorities in Engineering(NACME) Google Applied Machine Learning Intensive (AMLI) at the Illustrious Morgan State University
Developed by:
- Takiya Eastmond -
Morgan State University - Donald Davis -
Morgan State University - DeAndre Charity -
Morgan State University - Xavier Johnson -
Morgan State University
The project goal is to use Fibonacci’s sequence in a recurrent neural network to find the approximation of the following numbers in the sequence. We are using the Fibonacci sequence in a recurrent neural network to find the most accurate representation of the data. The sequence is used as a mathematical metaphor to see the bias and fairness that is used in today’s times as well as how it impacts people’s lives. The tools that we use to accomplish this task is the golden rule which is a quadratic formula that we implanted to teach the network a pattern. In which it accomplishes a divider for fairness. We are also using linear regression as a tool that will give us an exact representation of where the data needs to be so we know where our recurrent neural network needs to be.
Fibonacci’s sequence is the series of numbers where each number is the sum of the two preceding numbers, in his attempt to discover how many rabbits can be produced from a single pair of rabbits in one year under ideal conditions. The mathematical formula for Fibonacci’s sequence is Xn= Xn-1+Xn-2
The main thing that is valuable being able to take the recurrent neural network train it to understand a pattern and take that understanding to make accurate prediction
Our goals for this project is to build the most accurate and effective recurrent neural network with the intent to teach the model the Fibonacci’s sequence and construct a recurring neural network model to predict Fibonacci's sequence based on a predetermined sequence of data.
Using a Recurrent neural network, we measure the accuracy of prediction of the next numbers of the Fibonacci’s sequence. We then build a different model by using more of a LSTM (Long short-term memory) with control over the learning rate and Decay rate. Then based on these two models we build off those with additional models that we tweak to test if we get closer and more accurate results.
“The Fibonacci Sequence.” Imagination Station, www.imaginationstationtoledo.org/about/blog/the-fibonacci-sequence. Accessed 15 July 2022.
Velasquez, Robert. “What Is The Fibonacci Sequence? And How It Applies To Agile Development.” eLearning Industry, 12 May 2021, elearningindustry.com/fibonacci-sequence-what-is-and-how-applies-agile-development.
- Build a model that will be able to train a sequence of numbers
- Test a sequence to see how accurate and efficient the model is
- Build another model that will be able to train a sequence of numbers
- Test a sequence to see how accurate and efficient the new model is
- Compare and contrast the two models and build a third model with different hyperparameters that we received from the previous models
- Determine the most accurate
Please feel free to contact
Takiya Eastmond : [email protected]
Donald Davis : [email protected]
DeAndre Charity : [email protected]
Xavier Johnson : [email protected]