Each written algorithm is based on a particular problem but can be easily applied to other problems of the same level of complexity. Besides, all algorithms work for datasets of any size. Algorithms:
- Linear regression was aimed at predicting a) house prices given historical data on different houses in the area; b) revenue of a food company given their current locations and their historical populations and revenues
- Logistic regression was aimed at predicting a) 2-factor QA test results for microchips coming from the assembly line; b) student admission to a university given a student's ongoing marks
- Artificial Neural Network was aimed at recognizing hand-written numbers from 0 to 9 using 3-layer NN with feedforward propagation. File ex3_nn.m is the main for the algorithm. The folder also contains alternative approach of solving the problem through one-vs-all logistic classification which, however, produces only liner hypothesis and has a lower accuracy compared to NN.