Sihongy1 - Machine Learning (Level 1) Pathway

Certificates Summer 2021 Certificates Machine Learning Pipeline (NLP) (CE) 2021
1

Technical Area: Basic machine Learning knowledge Basic natural Language Processing knowledge Web scraping and web crawing

Tools: Beautiful Soap Selenium Pytorch Git Github Trello BERT

Soft Skills: Project management Coding skill

Achievements: Pracitce with Beautiful Soap and Selenium Know the basic of ML and NLP Tried web scraping

Goal for the next week: Create Trello for the team and manage the team with tasks

Tasks: Went over the instructions Watched the webinars in the resources Got familiar with what ML is Learned what NLP is, how it is used nowadays

2

Here is the link for my self-assessment 2: Machine Learning - Level 1 Module 2 - Sihong Yuan

3

Machine Learning Level 1 Module 2 (part2)

Technical Area:

  • Learned how to use Beautiful Soup and Selenium

  • Learned how to use webdriver

  • Learned how to Web Scrape and save data into a csv file

  • Learned how to do exploratory data analysis

Tools:

Beautiful Soap, Selenium, Jupyter Notebook, pandas, Tableau, chromedriver

Soft Skills:

  • Web scraping

  • Exploratory Data Analysis

Achievements:

  • Successfully scraped data from Flowster Forum

  • Cleaned the data by getting rid of the html

  • Transfered the data into lowercase

  • Removed stopwords of my data

Tasks Completed:

  • Scraped data from Flowster Forum

  • Explored my data after gathering it and analyzed it

  • Performed basic cleaning to remove html tags from data

  • Visualized my date

4

Here is the link for my self-assessment 3: Machine Learning - Level 1 Module 3 - Sihong Yuan

5

Machine Learning Level 1 Module 3 (part 2)

Technical Area

  • Vectorized my data using TF-IDF
  • Calculated a distance metric using cosine similarity
  • recommended 10 most similar posts based on the given post
  • Identified simple machine learning models and trained them on our data

Tools

  • sklearn
  • matplotlib
  • numpy

Soft Skills

  • Vectorized words using TF-IDF
  • Measured cosine similarity
  • Word embeddings
  • Simple classification model

Three achievement highlights

  • Transformed my data into word vectors using TF-IDF
  • Calculated cosine similarity
  • Recommended 10 most similar posts based on a given post

Goals for the upcoming week

  • start module 4
  • Prepare for the next presentation

hurdles

Had trouble to train the classification model

6

Machine Learning Level 1 Module 4

Technical Area:

  • BERT
  • Simple transformers
  • Logistic Regression

Tools:

  • Google Colab
  • Simple transformers
  • Sklearn

Soft Skills :

  • Simple transformers library
  • BERT model
  • Logistics Regression model

Achievements:

  • Used a pre-trained monolingual model to my dataset
  • Used DistilBERT to embeded the flowster dataset
  • Trained logistic regression model

Goals for the upcoming week:

  • Continue to finish module 4
7

Machine Learning Level 1 Module 4 （part 2)

Technical Area:

  • Learned how to use BERT and Roberta to update our word embeddings
  • Learned how to tested different recommenders to decide which one is the best
  • Learned how to built a classifier with group
  • Learned how to put all the work to a website

Tools:

  • Simple transformers
  • tokenizers
  • Google colab
  • Flask
  • Docker

Soft Skills:

  • How to measure the effectiveness of the recommenders
  • How to build a website using flask and Docker

Achievements:

  • Tested different kinds of recommenders and decided which one is the best
  • Put all our work into a web
  • Built a classifier and measured the effectiveness

Goals for the upcoming week:

  • Prepare for the final presentation