This is a progress report of my pledge to Siraj Raval's #100DaysofMLCode challenge. I will be updating here to links, activities of my work involving Machine Learning. Let's roll!
Link to my pledge: here
Work: Optimising power consumption of home appliances, study on Regularisation and Feature engineering
Thoughts: It was to fun to start off 100DaysOfMLCode. It was difficult to manipulate pandas using custom functions
Link to work: Github
Work:
- Optimising power consumption of home appliances, applied regression and plotted graphs of power consumption
- Created new dataset for power consumption by date, time, month
- Devices columns are added and new priority database of devices and weather is created
Thoughts: Because the data in randomly generated, the graphs look similar and symmetric in nature. Will change as real dataset is provided.
Link to work: Github
Work:
- Study on GridSearch CV
- Started the book Hands on Machine Learning with Tensorflow and Scikit
Thoughts: Machine Learning is more complicated than expected. The book that I am reading is pretty awesome!
Work:
- Coded custom algorithm to optimise power consumption of home appliances
- Plotted power consumption rates and minimised power consumption based on priority
Thoughts: Data generated is run across different trails and multiple regression algorithms are applied
Link to work: Github
Work:
- Generated new datasets and applied machine learning algorithms on the previous project
- Results are calculated and power consumption is compared
- Started the course, Tensorflow for Beginners
Thoughts: Tensorflow provides low level APIs and is much more complicated than expected. It also helps in creating custom machine learning models and algorithms
Link to work: Github
Work:
- Measured real data of Electrical Impedance Tomography
- New dataset is created for both male and female candidates
- Understood a couple of TensorFlow tutorials
Thoughts: EIT has high potential and can be used to classify affected human areas
Link to work: Github
Work:
- Measured impedance of people and logged real data
- Dataset is created and is pre-processed
Thoughts: As the real data is small and has high distribution, further investigation and hypothesis has to be made
Link to work: Github
Work:
- Analysed parameters for dataset
- Read through multiple research papers for measuring impedance from people
- Study on reccurent neural networks and convolutional neural networks
Thoughts: As the real data is small and has high distribution, further investigation and hypothesis has to be made
Link to work: Github Link to work: Github
Work:
- Participated in Target-HR Hackathon 12 hr hackathon
- Had to analyse and interpret results for a book dataset
- Completed the task given in time
Thoughts: One of the best hackathon I have ever been to. Kudos to Target HR for the great hackathon!
Link to work: Github
Work:
- Studied basics of Recurrent neural networks
- Trained a model on Convolutional Neural networks
Thoughts: Tensorflow GPU is much much faster than Tensorflow CPU
Link to work: Github
Work:
- Wrote documnetation for my project at work
- Prepared a presentation for the same
- Tested out some code to understand a couple of custom functions
Thoughts: Documentation is boring but important!
Link to work: Github
Work:
- Wrote report for the Home automation project
- Learnt about stock trading using Machine Learning
- Completed presentation given previously
Thoughts: Stock trading using Machine Learning has huge potential and opened my mind
Link to work: Github
Work:
- Prepared presentation for the Electrical Impedance Tomography Project
- Read about feature selection and feature engineering
- Started writing research paper on home automation devices
Thoughts: Writing research paper by using MS Word is a difficult skill!
Link to work: Github
Work:
- Completed two courses on datacamp
- Network analysis using NetworkX API
- Research paper on home automation devices
Thoughts: Network analysis is awesome!!
Link to work: Github
Work:
- Joining data using SQL on DataCamp
- Research paper on home automation devices
- Study on Neural Networks on Fast AI
Thoughts: Lots of work to do on Deep Learning
Link to work: Github
Work:
- Completed Research Paper
- Study on Regression evalution metrics
- Applied Regression evalution metrics to home database
Thoughts: Mean Absolute Error, Mean Squared Error, RMSE are intersting metrics
Link to work: [Github](www.github.com/nsudhanva/home
Work:
- Study on Matrices and Vectors
- Study on Neural Networks
Thoughts: Mathematics is really important for Machine Learning
Link to work: Github
Work:
- Completed Joining Data in SQL from DataCamp
- Completed Data Scientist using Python Track from DataCamp
Thoughts: DataCamp is a great place to get started for Machine Learning
Link to work: DataCamp
Work:
- Re-run tests and results for new values on the home automation project
- Completed research paper on home automation project
Thoughts: It's always great to write research papers
Link to work: GitHub
Work:
- Read about IBM Quantum computing and its applications
- Started working on previously created booking app
Thoughts: Quantum computing and AI can do amazing things
Link to work: GitHub
Work:
- Finishing up internship work
- Started watching Deep Learning AZ
Link to work: GitHub
Work:
- Finishing up internship work
- Finished report and presentation. Last day of work
Thoughts: Internship was awesome!
Link to work: GitHub
Work:
- Started predictive analytics course
- Continue watching Deep Learning AZ
- Started Affine Analytics Challenge ML
Thoughts: Predictive analytics is awesome
Link to work: GitHub
Work:
- Continue watching Deep Learning AZ
- Previous Code and Starting new project on Data Preprocessing
Link to work: GitHub
Work:
- Continue watching Deep Learning AZ
- Discuss possible machine learning projects from biotech department
Link to work: GitHub
Work:
- Continue watching Deep Learning AZ
- Study on AWS Machine Learning
Link to work: GitHub
Work:
- Continue watching Deep Learning AZ
- Study on LMS and Machine Learning for education
Link to work: GitHub
Work:
- Finished watching Deep Learning AZ
- Taught SQL to a class of 100
Link to work: GitHub
Work:
- Taught Advanced SQL to a class of 100
- Study on Recommendation Systems
Link to work: GitHub
Work:
- Study on Recommendation Systems
- Completed the course from Lynda on Recommendation systems - basics
Link to work: GitHub
Work:
- Study on Recommendation Systems
- Started recommendation systems using pandas course
- Tried to implement and relate recommendation systems to assignments
Link to work: GitHub
Work:
- Study on Recommendation Systems
- Data Analysis for one user and his assingment-submission timings
Link to work: GitHub
Work:
- Study on SVD
- Data Analysis for all user and his assingment-submission timings
Link to work: GitHub DCT
Work:
- Architect datapoints for DCT platform
- Implemented SVD to recommend related assignments
Link to work: GitHub DCT
Work:
- SQL to analyze and understand data
- Merge and Join data with sql and then pandas
Link to work: GitHub DCT
Work:
- Architect datapoints for DCT platform
- Analyze and recommend assignments for students based on another assignment
Link to work: GitHub DCT
Work:
- Study on Apriori ML algorithms
Link to work: GitHub DCT
Work:
- Built a recommendation system for recommending assignments to students using SVD
Link to work: GitHub DCT
Work:
- Built a recommendation system for recommending assignments to students using KNN
Link to work: GitHub DCT
Work:
- Study on AWS machine learning
- Study on recommendation systems using Implicit and Explicit feedback systems
Link to work: GitHub DCT
Work:
- Built a recommendation system using Implicit feedback system
- Used implicit library to improve sparse matrix computation
Link to work: GitHub DCT
Work:
- Built a recommendation system API to serve for DCT Platform
- API runs on Flask and is deployed to heroku
Link to work: GitHub DCT
Work:
- Study on Explicit recommendation systems
Link to work: GitHub DCT
Work:
- Study on Explicit recommendation systems
- Retrained model to match with assignments instead of submissions
Link to work: GitHub DCT
Work:
- Study on Implicit recommendation systems
- Deployed Flask model to Heroku and built and API around it
Link to work: GitHub DCT
Work:
- Used a new confidence metric system to evalute student behavior
- Deployed new model to Heroku and tested for correctness
Link to work: GitHub DCT
Work:
- API keys and authentication setup for code app
- Bug fixes and exception handling for the API
Link to work: GitHub DCT
Work:
- Setup API ends points to render json on recommendations
- Explore new options on recommendations for assignments
Link to work: GitHub DCT
Work:
- Reorder folders and made a common model folder
Link to work: GitHub DCT
Work:
- Wrote documentation on ML API
- Minor changes on model
Link to work: GitHub DCT
Work:
- Changes to documentation to the DCT ML API
- Added new algorithm to the request and number of assingments exception fix
Link to work: GitHub DCT
Work:
- Setting up AWS and SageMaker for Jupyter Notebooks
Link to work: GitHub DCT
Work:
- Changes to documentation to the DCT ML API
- Retrain the model and push the code to Heroku API
Link to work: GitHub DCT
Work:
- Gave a presentation on Recommendation Systems and API
- Started writing Research Paper on the same project
Link to work: GitHub DCT
Work:
- Gave a presentation on Recommendation Systems and API
- Started writing Research Paper on the same project
Link to work: GitHub DCT
Work:
- Gave a presentation on Recommendation Systems and API
- Started writing Research Paper on the same project
Link to work: GitHub DCT
Work:
- Gave a presentation on Recommendation Systems and API
- Started writing Research Paper on the same project
Link to work: GitHub DCT
Work:
- Gave a presentation on Recommendation Systems and API
- Started writing Research Paper on the same project
Link to work: GitHub DCT
Work:
- Gave a presentation on Recommendation Systems and API
- Started writing Research Paper on the same project
Link to work: GitHub DCT
Work:
- Gave a presentation on Recommendation Systems and API
- Started writing Research Paper on the same project
Link to work: GitHub DCT
Work:
- Gave a presentation on Recommendation Systems and API
- Started writing Research Paper on the same project
Link to work: GitHub DCT
Work:
- Gave a presentation on Recommendation Systems and API
- Started writing Research Paper on the same project
Link to work: GitHub DCT
Work:
- Gave a presentation on Recommendation Systems and API
- Started writing Research Paper on the same project
Link to work: GitHub DCT
Work:
- Gave a presentation on Recommendation Systems and API
- Started writing Research Paper on the same project
Link to work: GitHub DCT
Work:
- Gave a presentation on Recommendation Systems and API
- Started writing Research Paper on the same project
Link to work: GitHub DCT
Work:
- Started learning Node JS
- Understood the life cycle of a Javascript app
Link to work: GitHub DCT