Name: Ponkoj Shill
Type: User
Company: University of Nevada, Reno
Bio: Ph.D. Student in Computer Science.
Machine Learning Engineer.
Data Science Enthusiastic.
Location: Reno, Nevada, USA
Ponkoj Shill's Projects
This is the Ames Housing price dataset, used to predict the house price from test set using 80 different features.
Classifying several stages of Alzheimer's disease, a transfer learning based pretrained model(Restnet152v2) has been implemented. The dataset has been collected from kaggle that contains 6400 images with 4 classes of Brain images (mild demented, moderate demented, non-demented, very mild demented stages). At the initial stage, the pretrained model was used as the base model and few Conv2D, dropout, pooling and normalization layers were used on it. Moreover, the approach of training with the transfer learning based pretrained model performs better for this multi-class image classification.
Breast Cancer Prediction Using the Breast Cancer Wisconsin (Diagnostic) Database to create a classifier that can help diagnose patients.
This is a project to recognize cat using logistic regression with Neural Network concepts of backward and forward propagation from DeepLearning.AI.
Using pycoingecko library to fetch bitcoin price data from " https://www.coingecko.com/ " site. Used plotly Candlestick plot to visualize the trends as html format.
Here three datasets are used to analyze among countries regarding energy, power usage, GDP growth etc.
I have used Boston Housing Dataset, which is available at Sklearn API. Find out the outliers using box plot, scatter plot and clean them using Z score and IQR method.
Three datasets for these hypotheses testing such as Zillow research data site of housing data for the United States, a list of university and GDP of the United States in current dollars, in quarterly intervals are used.
Here, Matplotlib and Seaborn are used to show different kind of plotting visualization with python from scratch.
Building few types of deep learning models using Tensorflow from scratch.Defining the model requires that you first select the type of model that you need and then choose the architecture or network topology. This involves defining the layers of the model, configuring each layer with a number of nodes and activation function, and connecting the layers together into a cohesive model.
I used adult depression dataset from Health.Gov of California. After analysing the dataset, found some important inforamation regarding some most cases reason for the adults like age, sex, race, edication, income etc.
20800 train and 5200 test news dataset used to classify the fake and real news using Count Vectorizer and TF-IDF. Seven ML algorithms are applied to find the best model for the dataset.
A very bad and messy dataset which is in TSV format selected to show how to deal with and how to transform and scaling features for fit into a model.
I have collect all data FIFA dta from 2015 to 2020, then manipulates all dataset and visualize the dataset using plotly, matplotlib. Analyse the data to compare player position, growth, salary, market-value etc.
Trying different process of applying genetic algorithm.
Here, I used the Boston Crime data. Seaborn, folium, plotly are used to visualize the crime spot and analyze most crime occured area.
Create a public git repository called gitsub by forking from https://github.com/Znasif/GRAD-778Links to an external site. Clone the forked repository to your local machine Create a branch named: sub_movies