Ikabir - Machine Learning Pathway

Certificates June 2020 Certificates Machine Learning Pipeline (NLP) Self-Assessments
1

strong textConcise overview of things learned. Break it up into Technical Area, Tools, Soft Skills:**

Technical : I have gained more experience on python, specially pandas library. Have understood the basics of webscraping using selenium.

Tools : Have been introduced with Asana and GitHub.

Soft Skills :Collaboration with other team mates.

Three achievement highlights:

  1. Started collaborating in Github with team mates.
    2… Got introduced to Jupiter notebook and python pandas library and used it.
  2. Learned about webscraping and did some hands on work.

List of meetings/ training attended including social team events

Attended most team meetings and watched webinar videos related to webscraping and Github.

Goals for the upcoming week. Next self-assessment will be due on the following Tuesday 06/23

Clean the extracted data and use it. In the process, want to learn some new things on ML.

Detailed statement of tasks done. State each task, hurdles faced if any and how you solved the hurdle. You need to clearly mark whether the hurdles were solved with the help of training webinars, some help from project leads or significant help from project leads.

–Got introduced to and opened account on GSuite, Asana and Github.
Hurdle: Needed some help regarding opening the Gsuite account.

– Started using Jupiter notebook and python. Got familiar with pandas library. Did some elementary webscraping using selinium and pandas. Imported the data as pandas .csv data frame.
Hurdle: The whole concept of webscraping was new to me, and I am not a lot familiar with ML techniques. Got familiar with many things lately, and needed help from group members.

2

Concise overview of things learned. Break it up into Technical Area, Tools, Soft Skills:

Technical : Learned about BERT. Have learned a bit more about NLP, considering I am new on this.
Tools: DistilBERT
Soft skill: Communication.

achievement highlights:

Started to learn how to use DistilBERT to train a model

List of meetings/ training attended including social team events:

Meetings: Attended all team meetings on Zoom

Goals for the upcoming week:

Use DistilBERT for further analysis

Hurdles: Had no prior knowledge on BERT/NLP.

3

Full Session Assessment:

Concise overview of things learned. Break it up into Technical Area, Tools, Soft Skills:

Technical : 1) Web scraping, 2) Data organizing and cleaning, 3) Data feeding into BERT and doing classification

Tools :

1 ) Slack, 2) GitHub, 3) Python (pandas library) , 4) Asana, 5) DistilBERT

Soft Skills : Communication and team collaboration.

Three achievement highlights:

  1. Successfully scraped a category from Amazon forum website.
  2. Used pandas data frame to process data
  3. Implemented DistBert and analysis like logistic regression, random forest, etc.

List of meetings/training attended including social team events:

Team Meetings: All team meetings and final project presentation

Training Sessions/Webinars: attended Git Webinar. Watched the recordings of other webinars like NLP, Python training, etc.

Detailed statement of tasks done. State each task, hurdles faced if any and how you solved the hurdle. You need to clearly mark whether the hurdles were solved with the help of training webinars, some help from project leads or significant help from project leads.

Task 1: Set up Slack, Gsuit, Github, Asana

  • Task 1 Hurdles: Nothing much. Got help from the leads.

Task 2: Used Salenium and Beautiful soup for data Scraping.

  • Task 2 Hurdles: It was my first time scraping, so faced some trouble extracting informations out of the webpage webpage. Got lot of resources from the team 7 group and figured the task out.

Task 3: Data cleaning and processing

  • Task 3 Hurdles: Had no experience on NLP or BERT, but with teammate collaboration and help from the leads, successfully cleaned the data and processed it. It was a bit hard to grasp, but given enough time, was successful doing this task.

Task 4: Did classification algorithm by linear regression, random forest. Implemented DistBert

  • Task 4 Hurdles: Hurdle as well as major learning during this task was using several classification techniques like linear regression, random forest, neural network. Although neural network was not used, grasped some good idea about it while searching about classification algorithms. Used the classifiers as tools and got an idea about parameters to tune models.