Machine Learning level 1 Module 1 - Amath

Learned:

  • technical area : i discover beautiful soup which is a tool that help to scrape data from web pages, check the documentation and watch youtube to better understand the concept and how to use it. I have also a better understanding of git by watching the video and i finally build a model with logistic regression and naive bayes for sentiment analysis. I’m learning also NLP course(https://deep-learning-drizzle.github.io/) .
  • Tools : i install anaconda which content jupyter notebook, i use also beautiful soup and work a little bit with it. I get in touch with git which i think interesting for collaborative work. I use nltk for work tokenizing. I worked also with selenium, vscode, pytorch and tensorflow.
  • soft skills: there are many documentations in the web and i one the soft skill i learn is about how to use ressources and organise them in order to learn a specific task.For my self , i learn quickly by watching video through youtube for example.

Three achievement highlights

  • Git : After watching the video, i’m able to use git more efficiently , how to coordinate the collaborative work, how to keep track of any change on the project, version control.
  • beautiful soup : i install beautiful soup and go through the documentation in order to better understand the concept and watch also some youtube video about it.
  • dataset and build model : i use tweeter dataset from kaggle and build Naive Bayes and Logistic Regression models for sentiment analysis(good or bad). It give me a little refresh about basic ML model. I want to mention also that i build a class for the Logistic Regression and Naive Bayes models so i did it from scratch.

Tasks completed:

  • install the low level technical requirements : vscode, beautiful soupe, selenium, pytorch and tensorflow
  • I watch also the webinar and i think now i got the necessary tools we will need for the project.
  • I watch many ML course to get familiar with ML project workflow : upload the dataset - data preprocessing - split the data into train and test - model selection - train the model on training data- test the model on test data - improve the accuracy
  • I watch a video about the theory behind NLP(word embedding, tokenizing, transformers, LSTM, etc) and i finally build from scratch a logistic regression and naive bayes models for sentiment analysis (good or bad tweet).