Team for Click Prediction




I am creating a team for the Click Prediction Hackathon. Is anyone interested to be a member?
I have participated in IIM Bangalore sponsored by Amex hackathon.

Looking forward to join for a team.


I will join with you,even I’m looking for a team


I’m trying to understand what needs to be predicted first, may be relevant later.


I am trying to download the data - train and test data. It goes to the Google Drive and then shows the as 463 MB and as 1.1 GB. Is this true? I also tried using command level - wget like this:

wget -o destination_filename “

as well as, wget -o destination_filename "“

It downloads a file sized 300 kb but it corrupted and doesn’t open.

Can anyone suggest how to download the data file.

Thanks and Regards,


Hi @gops75,

Refer to this link


Hi even I m looking for a team member. Anyone interested to join?


Hi I am interested in joining team


I am also Interested yo be in team


I am also interested in join you team! can we get in touch with weixin or somethong else.


I’d like to be one.


At first look, it seemed only ConversionStatus has to be predicted which seemed like a classification problem. On a second look, I could also see ConversionPayOut which looks like a regression problem. What do you think?


I think it could be a Regression problem as in the problem statement it was given what Publisher, Network and Advertiser do with the commission part and in fact, the problem can also be revolved around in such a way that which publisherID and advertiser campaign creating max and min revenue.

What about your guess?


Sounds interesting, though they have explicitly mentioned they need a model to predict conversion on clicks. What I feel is instead of it being a regression problem, this is a classification problem.

The column ConversionPayOut appears to be a promising feature since the revenue will only be generated when there is a conversion and the magnitude revenue is a measure of how good that campaign was? Look at the following


After looking at these values this is a case of classification as only 633 got some numeric value and 4 fraud values out of false conversion rate.

That means a very low conversion rate.

Classification problem it is, then.

Thanks for this info as I took a guess reading problem statement only.


@mohdsanadzakirizvi,after cross-checking with the problem statement and data description, I would say it is a regression based problem as RMSE being an evalution metric is mostly used for regression and most of the times dep variable is placed at the end and so is the case with the data description used by AV.

This is my personal opinion


I’m actually new to data analytics and just completed an online course on data analytics with R. So I’m requesting if you can give me chance to be part of this team so that I can gain some realtime experience.


Hi Ashish

iam interested , but I learnt so far only Introduction to R ,Introduction to Python from Datacamp and have some SAS knowledge .please advise.


@Sekhar2017 I too am a student but you can practice from the below link where problem statement along with the data and code explaination is explicitly mentioned.


Me too Interested in making a team…