Gleb Koptev's repositories
average-salary-prediction
In this IPython notebook was implemented time series analysis on dataset of average salary in Russia.
Facial-key-points-regression
Facial key points prediction using Regression CNN with STN implemented on PyTorch.
oledb-csharp-to-ms-access
University task devoted to creation of console application which has access to MS Access database using OLEDB and implements database query
bagging-and-random-forest
The main purpose of the experiment represented in this notebook is to show how Decision tree and Random forest classifier algorithms depend from different hyperparameters and some approaches.
banners-placing
In this notebook we perform cluster analysis on Foursquare dataset to find the most optimal locations to place banners of Carnival Cruise Line company.
data-analysis-in-credit-scoring-task
This preliminary data analysis purpose is to discover dependencies and correlations between target feature and another features in dataset. Comments are on Russian.
endjouy
Website designed using Bootstrap and JS.
feedback-tonality-classifier
Feedback tonality classifier based on TfidfVectorizer and LinearSVC, trained on russian language feedback on smartphones. Interactive demo is attached too
BSU-2Degree-Programming
Laboratories
central-limit-theorem
In this IPython notebook I have conducted an experiment, which confirmed CLT on Wigner semicircle distribution example.
drinks
University project dedicated to polymorphism and inheritance
ISO-Labs
University studies :)
labyrinth
University task, which requires breadth-first search algorithm. The idea is concluded in the searching for the shortest path in matrix filled with 1 and 0 (available path and wall) between two points. My solution based on the Lee algorithm. So, I write 1 and 0 in my matrix as -2 and -1, which helped me to calculate and write to the matrix distance from the start position to the current point
linked-list
Simple linked list designed by me on C before my studying on FAMCS has started. Compiled on GCC, using Cygwin
linreg-height-prediction
Task from Yandex & MIPT Machine learning and Data Analysis courses. Consists in training linear model on ready dataset of heigths and weights with linear regression algorithm.
linreg-stochastic-gradient-descent
IPython notebook with Yandex & MIPT ML and DA Courses task. There is an implementation of linear regression to the dataset of company incomes and expenditures on advertising of different types. The main purpose is to build linear model which predict income according to expenditures on TV, radio and newspaper advertising.
medical-diagnostical-system
Supervised learning system, which monitors patients condition, diagnoses diseases and makes decisions. Designed on C# with XML-serialization to collect data about patients and binary serialization to collect templates condition-desease
Operating-systems-BSU
Operating systems 2 course labs
rectangles
Baltic OI'2004 task which was assigned as an university project. Solution on C++
scikit-uplift
:exclamation: uplift modeling in scikit-learn style in python :snake:
taxi-plates-regex
University task dedicated to regular expressions
text-similarity
Code based on this IPython notebook will analyse a set of sentences and find numbers of sentences, which are most similar to the 0 sentence (numeration from 0). Presented sentences are transformed to vector space and then cosine distances are calculated.