Pinned Repositories
medicare-fraud-detection
A machine learning based application to help detect fraud among Medicare practitioners.
sandeshregmi's Repositories
sandeshregmi/AnApplicationForClassifyingDepression
Abstract Depression brings significant challenges to the overall global public health. Each day, millions of people suffered from depression and only a small fraction of them undergo proper treatments. In the past, doctors will diagnose a patient via a face to face session using the diagnostic criteria that determine depression such as the Depression DSM-5 Diagnostic Criteria. However, past research revealed that most patients would not seek help from doctors at the early stage of depression which results in a declination in their mental health condition. On the other hand, many people are using social media platforms to share their feelings on a daily basis. Since then, there have been many studies on using social media to predict mental and physical diseases such as studies about cardiac arrest (Bosley et al., 2013), Zika virus (Miller, Banerjee, Muppalla, Romine, & Sheth, 2017), prescription drug abuse (Coppersmith, Dredze, Harman, Hollingshead, & Mitchell, 2015) mental health (De Choudhury, Kiciman, Dredze, Coppersmith, & Kumar, 2016) and studies particularly about depressive behavior within an individual (Kiang, Anthony, Adrian, Sophie, & Siyue, 2015). This research particularly focuses on leveraging social media data for detecting depressive thoughts among social media users. In essence, this research incorporated text analysis that focuses on drawing insights from written communication in order to conclude whether a tweet is related to depressive thoughts. This research produced a web application that performs a real-time enhanced classification of tweets based on a domain-specific lexicon-based method, which utilizes an improved dictionary that consists of depressive and non-depressive words with their associated orientations to classify depressive tweets. Problem understanding or Business Understanding Depression is the main cause of disability worldwide (De Choudhury et al., 2013). Statistically, an estimation of nearly 300 million people around the world suffers from depression. Shen et al (2017) mentioned that approximately 70% of people with early stages of depression would not consult a clinical psychologist. Many people are utilizing social media sites like Facebook and Instagram to disclose their feelings. This research persists the hypothesis that there are similarities between the mental state of an individual and the sentiment of their tweets and investigated the potentiality of social media (like twitter) as a data source for classifying depression among individuals.
sandeshregmi/b2b_bid_pricing_cm
end to end predictive modeling framework leveraging multiple ML techniques for bid pricing
sandeshregmi/BERT-pytorch
Google AI 2018 BERT pytorch implementation
sandeshregmi/Book_on_Python_Algorithms_and_Data_Structure
🪐 Book on Python, Algorithms, and Data Structures. 🪐
sandeshregmi/caret
caret (Classification And Regression Training) R package that contains misc functions for training and plotting classification and regression models
sandeshregmi/ClinicalBERT-Deep-Learning--Predicting-Hospital-Readmission-Using-Transformer
Blog post on Medium
sandeshregmi/CMS-Fraud-Detection
Fraud detection algorithm using Autoencoders and Stacked Autoencoders to detect fraudulent physicians in CMS Part B claims data
sandeshregmi/coronavirus_visualization_and_prediction
This repository tracks the spread of the novel coronavirus, also known as SARS-CoV-2. It is a contagious respiratory virus that first started in Wuhan in December 2019. On 2/11/2020, the disease is officially named COVID-19 by the World Health Organization.
sandeshregmi/Covid-19-data-science
Welcome to Glacier Data Project. A post-wuhan2020 project for data science
sandeshregmi/COVID-GA-model
Code for fitting and simulating a stochastic, mechanistic model of COVID-19 transmission in Georgia.
sandeshregmi/data_mine_journal
In 2003, I began my practice of journaling: writing down my thoughts, plans, fears, and introspections into a simple text file. Since then, I've written 631,597 words into 787 total journal entries, spanning 17 years. Journaling is an invaluable tool to better understand one's own mind: to analyze decision making patterns, identify trends, and find purpose in this confusing world. Over the years, I've conducted plenty of ad-hoc deep-dives into my journal. For example, when recovering from an injury, I used my journal to identify triggers that caused aggravation and plotted those over time to better understand how my recovery process was being shaped by my behavior. What if there was a way to do this type of analysis programmatically? Luckily, there is - it's called Natural Language Processing - and there are plenty of libraries that allow us to parse and analyze vast amounts of text. We will use TextBlob, a popular python NLP library built on top of NLKT and Pattern, to calculate the following: Descriptive statistics: words per year, entries per year, overall words. Most common words using word clouds and N-gram analysis (combinations of words) Parts of Speech tagging (Noun, verb, adjective). Sentiment polarity analysis - how has the tone (positive/negative) changed over time?
sandeshregmi/deep-learning-with-python-notebooks
Jupyter notebooks for the code samples of the book "Deep Learning with Python"
sandeshregmi/epi-mcmc
Scripts for epidemiological modeling of an epidemic
sandeshregmi/HealthDeterminants
Social Determinants of Health Visualization
sandeshregmi/intro-to-ml-tidy
Intro to Machine Learning with the Tidyverse
sandeshregmi/learn
RECON learn: a free, open platform for training material on epidemics analysis
sandeshregmi/Machine-Learning-with-Python
Python code for common Machine Learning Algorithms
sandeshregmi/mgm
Estimation of Time-Varying k-order Mixed Graphical Models
sandeshregmi/Natural-Language-Processing
NLP- Deep Learning, Gensim, Spacy, NLTK and Scikit Learn
sandeshregmi/nlp-in-python-tutorial
comparing stand up comedians using natural language processing
sandeshregmi/opioids
Analysis of opioid use in Texas
sandeshregmi/Optimization-of-Healthcare-Network-Facility-Staff
This repository contain a python notebook in which we create solution of allocating healthcare staff.
sandeshregmi/PythonDataScienceHandbook
Python Data Science Handbook: full text in Jupyter Notebooks
sandeshregmi/pytorch-beginner
pytorch tutorial for beginners
sandeshregmi/pytorch-nlp-notebooks
Learn how to use PyTorch to solve some common NLP problems with deep learning.
sandeshregmi/PyTorchZeroToAll
Simple PyTorch Tutorials Zero to ALL!
sandeshregmi/recommendation_system
sandeshregmi/sent-conv-torch
Text classification using a convolutional neural network.
sandeshregmi/shap-values
Shap values for model interpretation
sandeshregmi/textplot
Text Plots