There are few educational projects

ROOT

• Convolutional neural network for flower name predicting

  Making the functions to get the training and validation set from the Kaggle flower recognition task. In this notebook I've used several neural network structers using Tensorflow and Keras libraries. Also there is a Maxpooling impelementaion Link

• Richtracks prediction model

  Building a model that predicts whether the ride was made by a passenger or not based on data from incoming points. Score prediction measuring using raw prediction, data preprocessing with PCA method and configuration of hyperparameters Link

• Text classification to several categories using 20newsgroups dataset Link

• Review classification.

  The goal of the task is building a model that predicts type of movie review (negative or positive) using sklearn. We have two categories: "negative" and "positive", therefore 1's and 0's have been added to the target array. The folder contains two subfolders of .txt files divided into "negative" and "positive" reviews Link

• Spam, not spam sms prediction

  The aim of this task is to predict the category of the sms (spam or not spam) using logistic regression. Also we need to implement cross validation using GridSearchCV and finally compare scores (precision, recall, accurracy) of models. Link

• News parsing&clustering

  Parsing news of various topics from the https://iz.ru website. Finding optimal clusters number of dataset using elbow method. For each number of clusters value we will initialize K-means and use the inertia attribute to identify the sum of square distances of samples to the nearest cluster centre. To visualize, we’ll plot the features in a 2D space. As we know the dimension of features that we obtained from TFIDFVectorizer is quite large ( > 10,000), we need to reduce the dimension before we can plot. For this, we’ll use PCA and UMAP to transform our high dimensional features into 2 dimensions Link

Education

• Leaf classification

  Task is based upon Kaggle leaf classification competition. In my notebook I tried to use StratifiedShuffleSplit and visualize some important features. Also there is an implementation on PCA in order to reduce dimensions. In the end I tried to make a prediction using several classificators such as NearestCentroid, MultinomialNB and LogisticRegression. In addition there is a Random Forest implementation with GridSearchCV hupertuning.Link

• VK social network parsing

  Link

• Image detecting using сonvolutional neural networks.

  The idea behind this figure is to show, that such neural network configuration is identical with a 2D convolution operation and weights are just filters (also called kernels, convolution matrices, or masks) Link

• Image clustering.

  Clustering is a technique that helps in grouping similar items together based on particular attributes. We are going to apply K-means clustering to the image with 6-7 colors Link

• Moving object detecting using OpenCV

  At first we detect moving object then draw contours and fill with mean color between mask and source video frame Link

• Vandermonde matrix implementation

  Implementing vandermonde matrix without using numpy.vandermonde() Link

• Marathon simulation

  Implementation of a marathon simulation where the speed of runners is determined using the distribution law (exponential distribution, normal, Poisson, Bernoulli) Link

• Advertising search on Avito webpage

  In this task, we have to write a function that parses Avito webpage (we will only consider Moscow city). This function accepts two parameters. The first parameter is what we are looking for on the page. The second parameter is the number of the page to parse information from. You need to download the following information:

  • ad name;
  • ad url;
  • price;
  • subway station (if available), you need to carefully handle None or use the try-except construction;
  • how many meters from the subway station (if available).

  Link

Work

• Offline shop database.

  Данное приложение было написано на первом курсе обучения в рамках изучения дисциплины Python. Ознакомиться с руководством пользователя можно тут Link