Snehal's Projects
Recommender System is an information filtering tool that seeks to predict which product a user will like, and based on that, recommends a few products to the users. For example, Amazon can recommend new shopping items to buy, Netflix can recommend new movies to watch, and Google can recommend news that a user might be interested in. The two widely used approaches for building a recommender system are the content-based filtering (CBF) and collaborative filtering (CF).
Machine Learning University: Accelerated Computer Vision Class
Machine Learning University: Accelerated Natural Language Processing Class
Machine Learning University: Accelerated Tabular Data Class
Machine Learning University: Decision Trees and Ensemble Methods
Interactive Data Visualization in the browser, from Python
Interactive Web Plotting with Bokeh in IPython notebook
This project follows the traditional techniques like the Bag of Words and tf-idf to represent words in a corpus in a numeric format for multilabel classification.
Udacity ML
Udacity ML
In this article we will walk through the steps involved in collaborating over vcs to version control and proof read their codes.
This project aims to visualise covid spread in UK using a python visualisation package Bokeh.
Code for the online course "Deployment of Machine Learning Models"
This contains projects based on Algorithmic Marketing like Marketing Mix Modeling, Attribution Modeling & Budget Optimization, RFM Analysis, Customer Segmentation, Recommendation Systems, and Social Media Analytics
Driving lessons made simpler. Custom scheduling API built with Python.
PyTorch version of the paper 'Enhanced Deep Residual Networks for Single Image Super-Resolution' (CVPRW 2017)
Tensorflow implementation of "Show, Attend and Tell: Neural Image Caption Generation with Visual Attention"
In this project, we will understand how a machine reads and processes image for machine learning models. We will look into the underlying data structure of an image, packages used in python for image processing, convert the images into numpy arrays, split the dataset into train and test and end this part of the series by converting the datasets into tensors for deep learning models.
NVIDIA Merlin is an open source library providing end-to-end GPU-accelerated recommender systems, from feature engineering and preprocessing to training deep learning models and running inference in production.
The purpose of this study is to look at the distribution of ratings, movie and users over time, impact of user mood on average rating score and average rating score of genre over time. The analysis is divided into 4 di↵erent 5-year batches to run analysis on sections of data. It was found, the growth, trend and level are stable after the first 5 periods (i.e. after the year 2000). With frequency of rating showing high correlation to new movies and users added, trend for rating over time shows combining e↵ect of growth in user and movie base . Further, weekday-weekend analysis show most of the ratings (approx.70%) are happening over the weekdays. For average rating score, a notable observation is, the shift in the rating pattern for the last batch(latest batch, 2011-2015). In this batch approximately 50% of the rating scores are average and the 25% each for poor and high rating scores in comparison to the other batches where it was 80-20 between average and high/poor rating scores. In the genre analysis it was found 9.4% times users rated genre below 3, 17.5% times for high and 70% times average.
Spam detection is one of the major applications of Machine Learning in the interwebs today. Most of the email service providers have spam detection built in to automatically classify such mails as 'Junk Mail'.
Named-entity recognition (NER) (also known as entity identification, entity chunking and entity extraction) is a subtask of information extraction that seeks to locate and classify named entities mentioned in unstructured text into pre-defined categories such as person names, organizations, locations, medical codes, time expressions, quantities, monetary values, percentages, etc.
Udacity AWS ML
Tool for 1st Quality Control on Hydrographic Cruise Data.
Materials for DataScience.com LTV and Neural Nets Talks at PyData Seattle
Sagemaker provides tools to build, train, tune, deploy, and manage large-scale machine learning (ML) models, simpler. In this article, we will be looking at each of these steps.