Learn everything about Analytics

10 Machine Learning Algorithms Explained to an ‘Army Soldier’

, / 18


If you think deep, you’d realize the whole process of predictive modeling is a war. Ruthless war. Don’t you believe me?

Consider the data set as your opponent. Your knowledge of data mining & ml algorithms as your weapons. Your win will depend on the intuitive usage of your knowledge & strategy to get highest accuracy. So, how many times have you won this battle?

Here’s my short story.

I realized it last week. While travelling to south, I met a person whose daughter is a data scientist. That’s all he knew when I asked him about his kid’s profession. He didn’t get much time to spent with his daughter. He was a soldier.

He inquired about me too. His eyes glowed with ecstasy when he found out that I’m also a data scientist. He inquired further. Precisely, he was curious to know about ML Algorithms. It’s a lot being used in defense these days, he said.

This was a tough task for me. I didn’t know how to explain the technical stuff to a soldier. I found out a way. I knew he would surely understand things about battle, warfare, strategy. I explained him machine learning in these terms. In the end, he was convinced that his daughter has landed herself into a challenging job. He felt proud.

In this article, I’ve revealed that approach used to explain ML algorithms to the soldier. Though, there would be less soldiers who would read this, yet this idea promotes an interesting way of learning. This would surely help beginners who are struggling to understand these algorithms. I’ve also added images to help you visualize the situation and learn.

Note: The objective of this article is to help people learn machine learning in a fun and interesting way. You’ll also notice every algorithm has its own special situation where it is used.

This is how I started!

machine learning to army soldier


There are 3 Categories of Machine Learning Algorithms

Supervised Learning : You go on a war and don’t stop until you kill all your enemies whoever comes your way. It include algorithms such as Linear Regression, Logistic Regression, Decision Tree, Random Forest etc.

Unsupervised Learning : Your rival has challenged you for a war. Now, you decide after assessing your strengths and weakness, whether to accept the challenge or surrender. It include algorithms such as k-means, apriori etc

Reinforcement Learning : You’ve accepted the challenge. The war has begun. After every hour, you are accessing your position in the battle. Are you loosing more men? Is the opponent dominating? And, accordingly you decide where to continue or surrender. It includes algorithm such as markov decision process.


These are commonly used Machine Learning Algorithms

These algorithms are nothing but the weapons used to fight data sets. Make sure you learn them well to build your armory. Look, I’ve 10 weapons in my armory. How many have you got?

Let’s see my way of explaining these algorithms to the soldier.


1. Linear Regression

Linear Regression is completely taking over your opposition. Once you’ve entered the battle ground, don’t look back. Only take rest, once you’ve satisfactorily killed all your enemies.



2. Logistic Regression

Logistic Regression is taking simplistic assumptions and then deciding whether to go on a war or not.

In this photo dated July, 8 2013 Ragnar, front, played by actor Travis Fimmel, a leading character from television show Vikings, History’s brooding and brutal drama about the 8th-century Nordic warrior Ragnar Lothbrok. After a six-month shoot in Ireland, season two debuts Thursday night sporting a bigger scale, more confident pace and stronger entertainment than last year’s uneven, at-times plodding inaugural run.(AP Photo/Bernard Walsh/History)


3. Tree Based Modeling

This includes Decision Trees, Random Forest. Tree Based Modeling is simply Divide and Rule. You divide your opponents with some smart strategy and kill them.



4. Bayesian Modeling

Bayesian Modeling is considering your probability of winning in different battle base types such as air battle, land battle, water battle. And, accordingly take the overall decision of battle.



5. Support Vector Machines

SVM is drawing out territory and boundary where you’ve advantage on field base. For example, your soldiers might be adept to fighting in specific areas such as desert / mountain / fields and declare the war accordingly.



6. kNN (k Nearest Neighbour)

kNN is checking past outcomes and mapping accordingly. Evaluate your performance in past battles, contemplate on your weak and strong areas and prepare for the next battle accordingly.

U.S. Marines with the 13th Marine Expeditionary Unit (MEU), Alpha Company 1/4, receive a briefing before a platoon assault exercise at Arta Range, Djibouti, Feb. 10, 2014. The 13th MEU is deployed with the Boxer Amphibious Ready Group as a theater reserve and crisis response force throughout the U.S. 5th Fleet area of responsibility.(U.S. Air Force photo by Staff Sgt. Erik Cardenas/Released)


7. k-means

k-means is building up alliances (groups) with provinces which share the same philosophy, goals and motives. The idea is to become more powerful than ever before.



8. Neural Network and Perceptrons

Neural Network is every soldier in your army decides whom to fight. Imagine, the situation arising when soldiers run towards enemy soldiers and instantly decide which soldier to fight. Naturally, every soldier would prefer to fight a weaker enemy soldier (watch and analyze) so that he can kill as many quickly. Overall, this strategy would impact the performance of whole army.



9. Ensemble Modeling

Ensemble Modeling is when an army consists of men skilled in various combats such as archery, knife fighting, swordsmen, shooting etc which shares a common motive of ‘victory over enemies’.  All these men together would result in a formidable army.



10. Anomaly Detection

Anomaly Detection is checking for unusual patterns in your own army. You may have a secret agent among your soldier brothers. Keep your periodic checks on.



End Notes

I hope now you understand these algorithms well. This is how I helped that soldier to understand these algorithms. I believe, the most difficult things in nature, do have a simple explanation to offer. We just need to start relating this around us. As I just did. People new to machine learning, would now find them easier to remember and apply these algorithms.

In this article, I’ve explained machine learning algorithms to a soldier in terms of war, battle, and strategy. Do you find watching battle, wars interesting? If yes, you would surely find an interest here.

Did you find it useful ? Do share your opinions / suggestions in the comments section below. I’d love to talk.

If you like what you just read & want to continue your analytics learning, subscribe to our emailsfollow us on twitter or like our facebook page.


  • Mike S says:

    I recommend “The Master Algorithm” as a very readable and entertaining introduction to ML.

  • Jatin says:

    This was a really creative post.

  • Soumya says:

    Very well explained.

    It is said that if one can describe some complex things to a layman in simpler terms, then he / she has understood the subject a lot.

    Thanks for posting this beautiful article.

  • Santosh says:

    Nice explain…!!!
    Thank You….for u r efforts…

  • Harjit Chatha says:

    Dear Mr Saraswat
    I am a serving captain in the Indian Navy. These days I am trying to learn the nuances of Data Analytics. I am currently trying to get a hang of R. While looking for some web based modules to learn the language I chanced upon your blog. I have already subscribed to it.

    The present article “Machine Learning Algorithms Explained to an ‘Army Soldier” is very well written and has helped me better my understanding of the technical concepts.
    Thanks a lot.
    Eagerly waiting for your next blog !!!

  • Muhammad says:

    Sorry didn’t like it, it’s way too removed from reality with a little bit of truths here and there, but not even close

  • Cdr Abhijit Ghatak says:

    Excellent article; and by the way, a lot of soldiers, like me, will read and appreciate your article.

    I have a MS in Data Analytics (US) and presently pursuing my doctoral degree (while working for a EU project) in the area of Machine Learning and Pattern Recognition, in the Netherlands.

    Thanks Manish, I love reading your posts.

    – Abhijit

  • Johnny Johnson says:

    Nice! Makes me wonder. Would the layman (average business person) find value in the tools, packages, and techniques if they were to read something like this and drill down through it for more appropriate wording of the tools, packages, etc…. that would help them easily get to the meat of how it works and use it themselves.

  • Jakir Hossain says:

    This article will reduce fear about ML. Thanks.

  • Sunil Maurya says:

    Thanks for explaining the concept of ML in easy way.

  • Somnath says:

    In few minute you explain a lot.

  • Ram says:

    How about this?

    Here is a layman’s example of Predictive Modeling. I think Army Soldiers will have no difficulty in understanding this..

    -On Monday, it rained and it was cloudy, and we did not play outdoor.
    -On Tuesday, it did not rain and it was sunny, and we played outdoor.
    -On Wednesday, it did not rain and it was cloudy, and we played outdoor.
    -Thursday’s forecast is rain and cloud, and we predict ‘No outdoor games”.

  • Prof Ravi says:

    Well done Mr. Manish in particular and Mr. Kunal In general !!

    As posted above, making complex things simple is a difficult art in teaching and I guess this article achieves that to a great extent. Kudos for the effort. I love this portal very much.
    However, I just have a word of caution- It might be tempting to follow this approach in order to get more and more people hooked on, which is also important. But, it should not sound too unrealistically simplistic because novices prefer to be on the periphery of the subject wrongly believing that they already became ‘true’ data scientists. As you are aware, it is not that simple. I hope Mr. Kunal’s objective is to make serious/novice data scientists aware of a different approach to DS, which I liked very much. After all, all of us want ‘full-baked’ data scientists and not ”half-baked/quarter-baked” ones. Sorry for being didactic.

Leave A Reply

Your email address will not be published.

Join world’s fastest growing Analytics Community
Receive awesome tips, guides, infographics and become expert at: