Scribd
Upload
14 views3 pages

More About Planning A Machine Learning Project

The document discusses the importance of the Plan stage of the PACE workflow for machine learning projects. In the Plan stage, you define the problem, determine what type of machine learning model is needed to solve it (supervised vs. unsupervised, regression vs. classification), and verify that you have the necessary data and tools. Key aspects to consider include the goal of the project, the type of variables in the data, and whether the data needs preprocessing. The plan provides guidance but is not fixed - it can be reassessed as the project progresses.

Uploaded by

rayed786
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
14 views3 pages

More About Planning A Machine Learning Project

The document discusses the importance of the Plan stage of the PACE workflow for machine learning projects. In the Plan stage, you define the problem, determine what type of machine learning model is needed to solve it (supervised vs. unsupervised, regression vs. classification), and verify that you have the necessary data and tools. Key aspects to consider include the goal of the project, the type of variables in the data, and whether the data needs preprocessing. The plan provides guidance but is not fixed - it can be reassessed as the project progresses.

Uploaded by

rayed786
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
You are on page 1/ 3

More about planning a machine learning

project
The PACE workflow is something that can be used to keep the most experienced data professionals
on track in their projects. In this reading, you will learn more about the Plan stage of PACE and the
things that must be considered and determined to ensure a smooth and successful model
development process.

The Plan Stage


The PACE workflow is something that you can use to keep you on track, no matter the project you’re
working on. Each step is important to get to your final product. However, like many things, the most
important part is setting up the foundations of your project - The Plan Stage.

The Plan Stage is the part of the process where you first start thinking about what the problem
actually is, and what needs to be done to find a solution. You start to consider what tools you have
available to you, and how you’ll need to manipulate the dataset. Sometimes this can be as
straightforward as needing to create some visualizations for the data. Or, it can get as complex as
needing to make a predictive model using the dataset.

The plan that you create during this stage will be carried through the whole process, so it is
important to really make sure you’ve considered all the aspects and constraints of the project.
However, that isn’t to say that the plan you create must stay unchanging, you can absolutely
reassess as you progress. It is there to get you started heading in the right direction.

What should your Plan Include?


This section of the course focuses on machine learning algorithms, so we will use those types of
projects as the example here. However, you need to think about whether you need a model in the
first place! Many analytical tasks do not require the creation of a model, and you could spend time
creating something that is not necessary to what you’re trying to achieve.

Knowing What You Need For a Problem


The first thing to do when forming your plan is to consider the end goal. What exactly are you trying
to model, and what types of results from the model are needed? Something that can be determined
immediately is what type of machine learning model you’ll need. The two types that you’ve seen so
far are Supervised and Unsupervised models.

Supervised models are used to make predictions about unseen events. These types of models use
labeled data, and the model will use these labels and the predictor variables present to learn from
the dataset. And when given new data points, they’re able to make a prediction of the label. So, for
example, if you’re tasked with predicting rainfall amounts, you already know that you will need a
supervised learning model.

Unsupervised models, on the other hand, don’t really make predictions. They are used to discover
the natural structure of the data, finding relationships within unlabeled data. So, for example, if
you’re tasked with discovering relationships between customer habits and segment users, you know
you’ll need an unsupervised model.

Now, let’s go back to the rainfall example. Just from that problem statement alone, we know we
need a supervised learning model. However, not all supervised learning models are the same. The
two main types of supervised learning are Regression and Classification. There are different types of
regression models that you have practiced, with different models able to perform regression or
classification tasks.

Linear regression models are used when the result must be a continuous variable. As you have
learned, continuous variables are numerical values that can have an unlimited number of values
between the highest and lowest points of measurement. So if you need rainfall amounts in inches or
centimeters, you know a linear regression model is needed.

However, what if we don’t need exact rainfall amount predictions, but just whether or not it will rain
that day? This is where a classification model, such as a logistic regression model, would be more
appropriate. Classification models will deliver results as a categorical variable, where there is a finite
set of values that the variable can be. In this example, the model would only ever predict two results:
Will Rain or Won’t Rain.

Figuring out the tools you need


After you’ve determined what type of model you’re going to need, you must consider what you have
at your disposal to complete the project. Most importantly, you need to figure out if you have the data
you’ll need to build the model.

If your dataset only has one or two predictor variables, it probably will not produce a model that will
be useful. Or, if it has very few data points, the model’s performance will similarly suffer. On the
other hand, your dataset might be large and unwieldy, meaning that you’ll either need to clean it up
or cut it down to get it into a format that you can use to train the model. Having these issues means
that you’ll have to put in a little extra work to get it to usable form, or look elsewhere for data that will
be helpful to create the model.
Key Takeaways

 The PACE workflow for machine learning is very useful for planning out and solving data
driven problems.
 The Plan stage of PACE is one of the most important, setting you up for success throughout
the rest of the process
 In the Plan stage, you first consider the problem at hand and what will be needed to solve it
 You also verify that you have the tools and resources you need to solve the problem
 The Plan is not set in stone. It just serves as a foundational starting point for the rest of the
project

You might also like