In this course, five different projects were conducted, each of which concentrated on a particular concept of artificial intelligence. Their full description is available in the relative directory. Besides, a brief description of each project is included in the following:
Machine Learning: Designed a system to estimate house prices based on their attributes. The system was trained by a dataset from Kaggle website containing several house prices and their features. For this purpose, the dataset was pre-processed, several machine learning models were tested, and their pros and cons were pointed out. You can find the report here.
Designed a lung disease recognizer in the shape of a 3-layer neural network, trained by images of sample lung CT Scan records. The network had the ability to recognize the disease shown by input CT Scan images and classify them into three distinctive classes, each representing a specific lung disease. This project was conducted in two different phases. The first phase included designing a neural network from scratch, and the second contained designing the neural network with the Keras library.
IPYNB Notebook for the first phase, IPYNB Notebook for the second phase
Designed a system to estimate house prices based on their attributes. The system was trained by a dataset from Kaggle website containing several house prices and their features. For this purpose, the dataset was pre-processed, several machine learning models were tested, and their pros and cons were pointed out. The report (IPYNB Notebook) is available here.
Designed a comment classifier to detect positive and negative sentiment in users’ comments based on the conditional probabilities calculated by ”Bag of Words” technique. Files and more information about this project are available here.
Solved a problem using genetic algorithms inspired by nature and natural selection, in which genetically evolved solutions -known as chromosomes- are generated and evaluated in terms of closeness and similarity to the actual solution. In this problem, we were supposed to find the correct combination of six logic gates matched with a given truth table. Files and more information about this project are available here.
Designed a game simulating snake moves in a pre-defined table in order to execute informed and uninformed search algorithms (such as BFS, A-star, etc.) and compare their performance in terms of efficiency, admissibility and consistency. These search algorithms were used to find the shortest path for the snake to reach a final state. Files and more information about this project are available here.