vb100

AnomaliesDetection-with-TimeSeriesAnalysis

This is an experimental of anomalies detection by applying different approach to the problem. PCA component regularization method, K-Mean Clustering, SVM and Gausian Distribution models has been used to detect anomalies on time series data.

Deep-Learning---Recognize-Infected-Cells-by-Malaria--Adam--Sigmoid-

The project contain folders for labeled infected and non-infected cells by malaria. The code define an 4 layers ANN that learning to recognize infected cells by a given shuffled examples from working directory and provide a learning curve as output. For optimization Adam and MiniBatch methods have been applied (Tensorflow).

deploy-ml-mlflow-aws

Deploying Models to Production with Mlflow and AWS Sagemaker

Ensemble-Learning-Testing-Classifiers

Testing and Exploration of Ensemble Learning with poor dataset (66 obs). Results are provided at the end of the Notebook.

multivariate-lstm

Pneumonia-X-Rays-of-Human-Lungs-AI-project

This projects includes both back-end (Keras, Flask with dependencies) and front-end/deployement (HTML, CSS, JS) parts. The project perform X-rays of human lungs classifiaction using Convolutional Neural Networks with automated image pre-processing and internal procedures. Saved files of inspected models are saved on specific folders in project directory.

Tensorflow-ANN-TimeSeries-Forecasting-Weather

This notebook demonstrates how to implement Tensorflow & queries to forecasting for a very big dataset. ANN with 2 hidden layers has been used for ANN. The data shifting technique has been used to generate Train and Test (validation) sets.

Visualize-3D-MRI-Scans-Brain-case

The notebook introduces how to visualize 3D MRI data (Brain Scans case)

whisper_ai_finetune

Fine-tune WhisperAI model to your language

XgBoost-Evaluate-Data-Features-Machine-Learning

This code uses innovative XgBoost Machine learning algorithm that analyze data features and evalute it's importances in order to make data driven solutions in real estate. © Vytautas Bielinskas

vb100/Yahoo-Finance-Stock-Scrapper-Selenium

This Selenium based Scrapper gather historical prices of stocks that are defined in the external CSV file, collect all the price values with it's timestamp, then structuring data and write to Excel file with readable data structure where timestamp correspond to all stocks that has a price value at that day. All scrapping proccess takes even few day/nights because of Selenium speed and set delay time. Algorithm made by Python. © Vytautas Bielinskas 2018

vb100/Booking.com-Price-Scrapper-Selenium

This Python algorithm will gather data of prices for 365 next days for specific chosen hotel on Booking.com. There can be some bug for testing on different properties but for our target property is worked perfectly. (© Vytautas Bielinskas)

vb100/Booking.com-Data-Scrapper-Selenium

This Python made web scrapper gather all possible data (features) of properties in chosen area in ant place of the world. Due to dynamic content that is delivered by JavaScript scripts, Selenium was choosen as the best Python libary to gather such data. © Vytautas Bielinskas 2017

vb100/immobilienscout24.de-Scrapper

This Python based Web Scrapper gather all possible values of real estate properties features from www.immobilienscout24.de that is the biggest Real Estate web listing page in Germany Real Estate market. This scrapper gather data for rental properties. Suprising is that only BeautifulSoup was enough to build this scrapper. And magical thing - Geocoding by connection to Google Maps API is implementing inside this Python code! © Vytautas Bielinskas 2017-2018

vb100/XgBoost-machine-Learning-Template-including-Hyperparameter-Tuning

This Machine Learning algorithm [Python] evaluate data features and deliver it's importances with the best set of hyperparameters of XgBoost Classifier. It has been implementing by Hyperparameter Tuning. Easy adaption to any othet ML algorithm. © Vytautas Bielinskas

vb100/Zoopla-Scrapper-by-Selenium

This Python made web scrapper uses Selenium and BeatufulSoup to gather data of rental propertien from one of the biggest real estate listings website in UK. Just paste the search URL and click run. There are another scrapper for Sales and improved scrapper to collect some historical data. © Vytautas Bielinskas 2017

vb100/Automatic-Data-Parsing-Structuring-Grouping-Writer

This Python code read CSV file and automatically recongizes real estate properties, finds specific cells at Excel file where specific values should be written, make all Excel formating and building all neccesary formulas only by one click! Half day work done just in half a minute.

vb100/Immobilescout24-scraping-by-submarket-and-plot-it

This web scrapper gather all possible data (features) from immobilescout24.de for each selected district (let's call is as 'submarket') and save it to separate Excel file with unique timestamp. There are also data visualzation part in the code that can be switchen on/off. © Vytautas Bielinskas

vb100/K-NN-Machine-Learning-Neighbors

This K-NN Machine Learning algorithm take data from external source and testing Train and Test sets to get values of accuracies delivered by algorithm based on different number of n_neighbors (hyperparameter of KNeighborsClassifier). © Vytautas Bielinskas 2018

vb100/Real-Estate-Description-Evaluation-EDA-for-ML

This Jupyter Notebook file import scrapped data of real estate properties from www.zoopla.co.uk (see other resporities) and evaluate descriptions that under a property. Those descriptions are being scrapped during Jupyter Notebook implementation. Evaluations are assigned based on user specified Bag Of Words (BOW). It is classical Machine learning task, calling Natual Language Proceesing (NLP). Later Description evaluations lies on an array that is appended to whole dataset and used in ML algorithm later. There Random forest are using. © Vytautas Bielinskas 2018

vb100/Selenium-automation-data-gathering-and-writer

This Python code gathering numerical data from Stat. data website by Selenium, parse tables, structuring data and write all the values directly to Excel file including cell formating (OpenXlsX library).

vb100/Stock-Analysis-and-Visualization-Candles

This Python application read real stock data from Google API and represent graph on HTML page. Bokeh is used to build the graph. © Vytautas Bielinskas 2017