Meet The MachineHack Champions Who Cracked The ‘Power Plant Energy Output Prediction’ Hackathon

MachineHack efficiently concluded its eleventh installment of the weekend hackathon sequence final Monday. The Power Plant Energy Output Prediction hackathon supplied the contestants with a possibility to revisit the very fundamentals of information science and machine studying drawback. The hackathon was tremendously welcomed by knowledge science lovers with over 300 registrations and lively participation from near 200 practitioners.

Out of the 200 opponents, three topped our leaderboard. In this text, we’ll introduce you to the winners and describe the strategy they took to unravel the issue.

Out of the 200 opponents, three topped our leaderboard. In this text, we’ll introduce you to the winners and describe the strategy they took to unravel the issue.

#1| Devrup Banerjee

Although Devrup discovered python simply out of the sheer must automate the routine work and collect knowledge at scale, his actual enthusiasm and keenness for knowledge science sprouted in his second yr of MBA at Great Lakes Institute of Management, Gurgaon, whereas he was attending his advertising and marketing and retail analytics class. He realized that the true motivation behind studying all these algorithms was not about enhancing accuracy however to inform your consumer by how a lot you may promise to extend their bottom-line in the event that they have been to observe your precise given path. The topic modified his life. 

My faculty GreatLakes Institute Of Management has been instrumental in introducing me to this area. Till my commencement, I didn’t even know such a area existed and one may even take analytics as majors.  It was solely throughout my MBA years that I used to be launched to this extremely intriguing area. When I used to be younger I stored listening to individuals say observe your ardour, solely then you may obtain success, and eventually, once I was stored up at evening resulting from issues associated to datasets, the place perhaps I wasn’t in a position to calculate a sure metrics or attain a sure accuracy, I finally understood the which means of following of ardour. I’ve achieved some sort of success on this area albeit minor in comparison with the ocean that knowledge science is however then am solely getting began. I truthfully really feel our college and undergrad schooling system wants to vary to at the very least give college students the prospect to take a look on the full catalog of topics accessible that are a lot extra related professionally, for one to finally discover their actual ardour in schooling.


Machinehack is doing a great job organizing the hackathons and data science summits which not only provides exposure to budding data scientists and students like me but also helps in opening up networking opportunities with the very best in the domain. The practical approach always is better than theory and machine hack is one of the leaders when it comes to that” – Devrup shared his opinion about MachineHack

Approach To Solving The Problem 

Devrup explains his strategy briefly as follows:

The practice and check set contained a lot of duplicates. Now a few of these duplicates in check precisely matched these of practice. So I grouped the values in practice by imply of goal and manually imputed all the precise matched by this implies within the check. This alone gave me a rating of two.30 on the leaderboard. The remaining values within the check set have been predicted by a stacking mannequin between CatBoost and LightGBM. The automated CharacteristicTools was used for characteristic engineering objective and after some stable characteristic choice this gave me a superb rating.

#2| Akash PB

Akash did his bachelor’s in Statistics and Masters in Operations Research from IITB. From his faculty days, he was excited by programming algorithms from scratch and therefore got here to this area. On one positive day, he discovered a valuable gem on Coursera – Andrew Ng sir’s course, and after completion of his course, he felt just a little bit assured to use stuff in the true world. The firm he labored for gave him the chance to be part of a whole lot of initiatives that concerned machine studying and knowledge science and therefore he earned a real-life taste of theoretical information. Since March, he has been collaborating in MachineHack hackathons and from there I’m studying much more.

Machinehack is a nice place for me to experiment with things. Moreover, the interface is very nice now as compared to the previous one. The hackathons are nice and hopefully, we will see more challenging stuff in the future.”- Akash shared his opinion.

Approach To Solving The Problem 

He explains his strategy briefly as follows:

In this hackathon, knowledge was small and therefore one of the best answer could be from mixing solely. I used xgboost and blended Four fashions of xgboost with totally different parameters. The weights of mixing weren’t achieved manually. I made positive that the mannequin giving one of the best RMSE (least RMSE) in a fold will get the next weight and used exponentiation for that. Furthermore, I did some EDA and created a brand new characteristic by dividing ambient stress with Vacuum.

#3| Tamil Selvan

Selvan did freelance digital advertising and marketing throughout his faculty days with out understanding the time period itself. Because that was a time the place the time period digital advertising and marketing just isn’t within the hype.  In his closing yr, he acquired a possibility to market a restaurant in Chennai.

See Also

It was when he got here to know that automation helps the restaurant trade to make a greater expertise for purchasers in lots of elements. He researched loads, acquired love in direction of knowledge science machine studying, and AI. My constant need to be taught extra introduced me to pursue Data Science in Imarticus Learning, Chennai.

Now He is growing a no-code automated machine studying platform, the place one can do EDA, preprocessing, modeling, and provides reviews with out coding.

Machine hack is a wonderful platform for beginners. I really love to apply my theoretical knowledge on their hackathons conducting every weekend. Happy to be a part of the talented machine hack community”- Selvan shared his opinion.

Approach To Solving The Problem 

He explains his strategy briefly as follows:

No excessive skewness within the response variable, so no transformation is required. He created a brand new characteristic by dividing the Pressure variable with Vacuum. Another characteristic he created was a Dew level. It is the temperature to which air should be cooled to grow to be saturated with water vapor.

He used unsupervised studying to create labels, the place he used these labels as predictors for my mannequin. Yet, one other characteristic was created which is the distinction between Relative humidity and temperature. Finally, he used a 10-fold XG enhance with max depth and studying depth as parameters.

Provide your feedback under


If you liked this story, do be part of our Telegram Community.

Also, you may write for us and be one of many 500+ specialists who’ve contributed tales at AIM. Share your nominations here.

Anurag Upadhyaya

Experienced Data Scientist with a demonstrated historical past of working in Industrial IOT (IIOT), Industry 4.0, Power Systems and Manufacturing area.
I’ve expertise in designing strong options for numerous purchasers utilizing Machine Learning, Artificial Intelligence, and Deep Learning.
I’ve been instrumental in growing finish to finish options from scratch and deploying them independently at scale.


Please enter your comment!
Please enter your name here