DEVELOPING STATISTICS, DATA TREES AND DATASETS DESIGN TOOLS

​

Since the Third Wave Revolution began, data became one of the most popular notions. You can't escape the rise of big data and data science. This research is about machine learning as a design tool. A number of different algorithms were assigned in this machine-learning process. Supervised learning was mainly used with a set of labeled trading data. For every example, in the training data process, I had an input and an output object. In order for those supervised learning classifiers to know the result of each geometrical movement, I had to manually enter the preferred answers. ​I was mainly focused on the question of how to construct a computer program that automatically improves with experience. The goal was to present the key algorithms and theories that form the core of a machine learning shape generator. The project asks whether and how will the design process change with Artificial Intelligence techniques

​

​

COMPUTATIONAL INTELLIGENCE APPROACHES

EVALUATION VS. SIMULATION

​

Machine-learning algorithms generally fall into two classes, depending on whether the training data set includes the “correct” or the desired answer (supervised learning vs. unsupervised learning). Another branch of this research focuses on the field of neural networks, fuzzy systems, and evolutionary computation. The goal was to provide an in-depth understanding of the different techniques developed in each step of the project. There is a spectrum of philosophies about how closely AI should resemble and copy natural intelligence. In the end, I believe that properly written algorithms on conventional digital computers can efficiently produce the most important capabilities systems. Here, learning is accomplished by adjusting “performance” weight at each synapse, while modifying connections and strength.