When you’re reading an article, review, or book, extracting meaning is the point. But when you must read thousands and thousands of articles, reviews, or books extracting meaning becomes a bit of a challenge. It can be overwhelming to consider extracting and cataloging and tagging topics while reading, especially considering that some words can overlap topics. Sometimes words that can help classify a review as a five or one star are useless in telling you what you’re doing right or wrong.
When using machine learning to identify patterns in order to make predictions and your model yields an accuracy rate of 99% you are feeling really confident . Except… when in only 1% of the samples you’ve used to train your model reflects the one thing you’re trying to classify… the target. That 99% accuracy the model is touting is it’s ability to identify the samples that are not the target. Imbalance happens… and in this case.. despite your high accuracy… your model sucks.
I’ve said before the world is an imperfect place. Not all distributions are Gaussian, symmetrical and perfect in samples or populations. In machine learning, classification is a supervised learning appraoch in which the computer program learns from the data and make new observations or ‘classifications’ by using targets, or ‘classes’ that are assigned.
Hypothesis testing is a vital part of statistics. It essentially takes two statements and evaluates which the sample data best supports statistically. The results determine if there is significance. Real, measurable and quantifiable significance.
The first steps are sometimes the hardest. I found this to be true when starting the Data Science course at Flatiron School.
Like my classmates, I did most of the pre-work and was nervous about the interviews required to be accepted into the program. I was so excited when I was admitted. And then nervous. Having come from a background NOT coding or in math - my learning curve is very steep.
The first couple of weeks I felt completely overwhelmed. Like - really overwhelmed. Like I was sitting at the smart kids table and wondering how they let me do it. I felt like way in over my head. There was so much material and process being thrown at us daily. NOTHING made sense. Hoenstly - I wanted to quit. When I saw what was expected of us - even looking at the first project that was due - I became so anxious and had serious doubts.
I had talked to a copule of other students who were ahead of me who shared that when they began they felt the same, and in talking to my ed coach - she assured me that I wasn’t alone in these feelings. Being a bit of a data nerd, and perhaps I am a data scientist at heart - I began to plot how I felt daily on a scale from 0-5. 0 = quitting. 5 = YES! I made the right choice! I love this! I reasonded that if there were enough 0’s I would have made an informed decision using data.
Below is my chart. The data speaks for itself. My advice to anyone struggling. Keep showing up and give it an honest effort. Pay attention. Communicate - with your instructor, ed coach, cohort - stay connected. Learn strategies. Then see for yourself:
<img src=”https://raw.githubusercontent.com/andiosika/andiosika.github.io/master/img/DS%20Feels%20Image.png” ,width=”80%”,style=”text-align:center;” /img>