Vukašin Vasiljević's Projects
Analysis of prices in some of the largest supermarkets in the Republic of Serbia (mainly supermarkets in Belgrade). Supermarkets are: Maxi, Tempo, Idea, Dis, Univerexport, Lidl.
JavaProject
Classification metrics such as confusion matrix, accuracy score, recall, specificity, precision implemented in jupyter notebook
Solutions for Algorithms Part 1, on Coursera
Impact of COVID-19 pandemic on used cars market in Serbia, analysis of electronic market of used cars. Site polovniautomobili.com
Implementation of decision tree classification without EDA and other visualizations. Included simple tuning of hyperparameters with GridSearch and cross-validation with k-folds.
Calculating accuracy and roc-auc score with KFolds cross-validation on dataset using Common Workflow Language.
Given two sequences, find the length of longest subsequence present in both of them. A subsequence is a sequence that appears in the same relative order, but not necessarily contiguous. For example, “abc”, “abg”, “bdf”, “aeg”, ‘”acefg”, .. etc are subsequences of “abcdefg”. So a string of length n has 2^n different possible subsequences.
Car leasing calculator.
Machine learning web application that predicts the price of used cars based on car parameters. Furthermore, it incorporates a recommendation model for similar cars and it uses graphs to visualize price comparison between similar cars.
Python script that checks if certain change has been made to some web page.
Supplementary Material for “The mass-, period- and angular momentum dependences of the “moth-I” family of 3-body orbits”
Script that extracts information from car ads from website and collects them in mysql database for later use.
Working with DICOM files and segmenting aorta on lung pictures if one exists.
We analyze the world set of data for planar periodic three-body orbits with zero angular momentum that pass through a symmetrical collinear configuration (syzygy), and find correlations between scale-invariant parameters as the ratio m3 m of the central m3 and the equal m masses, the topological quantifier Lf (the number of letters in the free-group F2(a, b) word describing the topology of an orbit, as defined in Refs. [1, 5]), the scale-invariant kinematic variables, such as the scale-invariant period T|E| 3/2 , or the scale-invariant angular momentum L|E| 1/2 , on one hand, and the distribution of marginally stable orbits, on another. We find a distribution that falls within a compact non-zero measure domain with clearly defined, apparently smooth boundaries in the space spanned by the topological index Lf and the mass ratio m3 m . On this basis, we predict an increase of the number of linearly stable orbits, as the mass ratio m3 m falls below 0.5. We also predict the (approximate) domain of these orbits’ initial conditions. Some consequences for astronomy/astrophysics are briefly discussed.
Here's an example of convolutional neural network with UNET architecture that's used for segmentation of an lung X-ray medical images of COVID-19 patients.
This project is an attempt to create machine learning model for predicting car prices based on some key parameters that cars have; such as brand, model, mileage, motor volume, power, petrol and chassis.
This project tries to create fake data for users and use them with scraped and cleaned data of used cars.
The Wiki Game is a hypertextual game designed to work specifically with Wikipedia. Objective is to start on the randomly selected article and navigate to another pre-selected target article but this Python script does searching instead of you and provides path from start to destination article.
Creating neural network model for predicting type of animal based on given attributes