29 - Assignment1

You might also like

Download as docx, pdf, or txt
Download as docx, pdf, or txt
You are on page 1of 2

Purpose

In this forum, you will exchange ideas about how to model a specific machine learning problem.
Modeling the problem is the first step in any machine learning application. To model the
problem, you need to state 1) what type of machine learning problem it is, i.e., regression or
classification, 2) what the input of the machine learning model will be, i.e., its features, 3) what
its output will be, i.e., what the model aims to predict given the input, and 4) what dataset you
can use to train the model.

Instructions
The question you will answer in this forum is:
Imagine you are asked to build a machine learning model to detect the likelihood of a person
getting a heart attack. How would you model such a problem? To help you modeling this
problem, try to answer the following questions:
1. What type of problem is this? Is it a regression problem or a classification one?
2. What could be the features that the model could use to detect whether a person is
likely to get a heart attack or not?
3. What would be the output your model aims to predict? How would you represent
that output?
4. What dataset you would need to train your model?

Respond to this question by clicking on the post below [80 points].


After you submit your post, you will be able to view other posts (typically 15 minutes after you
post).
Comment briefly on at least two other posts by your classmates [20 points].

My answers:

1. What type of problem is this? Is it a regression problem or a classification one?

The likelihood of a person to getting a heart attack is a classification problem where the
machine learning model shall predict if a person with certain medical symptoms and/or
some characteristics (age, eating habits, weight, etc.) might get a heart attack or not.
2. What could be the features that the model could use to detect whether a person is likely
to get a heart attack or not?
Features that the model could use are:

1- Medical symptoms like: Chest Pain, Cold Sweat, Nausea, Discomfort etc.
2- Profile characteristics like: Age, weight, sex, eating habits, smoker, drinking habits,
etc.
3. What would be the output your model aims to predict? How would you represent that
output?
The output The output is typically a value representing the risk of getting a heart attack.
For instance if we use perceptron learning algorithm where we assign a weight to each
feature, a negative output of PLA would indicate a low risk of getting a heart attack and
positive output would indicate a high risk of getting a heart attack.

4. What dataset you would need to train your model?

The dataset should include: age, weight, sex, location, eating habits, smoking habits,
drinking habits, medical history, medical symptoms (Chest pain, Cold sweat, Nausea,
etc.)

You might also like