AboutI am a person who loves everything computer. Being seeing computers from the ye ol 486 days till now. Started with Dos, then Mac then windows and now linux. Interested in creative aspects of pgramming. might write a book on programming one day.
Joined devRant on 5/30/2018
Do all the things like ++ or -- rants, post your own rants, comment on others' rants and build your customized dev avatarSign Up
From the creators of devRant, Pipeless lets you power real-time personalized recommendations and activity feeds using a simple APILearn More
@singh029 is that a joke or a serious reference for the project?
@oudalally thank you so mucha again. Will dig in. I was thinking to break the project into smaller feasible units. Guess will try out with something small and workable.
Will post the updates. Thanks again.
@adempus any particular ide or writer app u wanna develop this? Coz i was also thinking on similar lines for sublime text 2.0.2.
@Joehaz start with some online tutorials for java or csharp if u cant find one. Sky is the limit. All the best.
@oudalally wow! Thank you so much! You read my mind. You made my day!
I know and understand the helplessness at not being able to solve this challenge. But somehow my mind tells me it can be done.
I am remembered of alan turing of making the bombe machine in using it to predict the position of uboats. And this is what inspired me to think and ask myself " gee, those folks at the institutes know atleast 60% of the questions coming in the exam. And it takes then an year to do so. What if i can do that in days or seconds by a computer? All this computing power and no prediction. Common!."....
Looks like i need to dig deeper. What about deep learning?
Thank you so much once again. :)
Next i learned .net i.e csharp in detail from a mentor. He taight me good, but learned oops in detail.
Then i touched nodejs and thats when i discovered it a lot more.
Learn java or csharp as its easy for oops. Js is hard for learnig oops in the beginner level.
@gitreflog i fully agree with u. But that said i am not looking at 100 % accuracy. I am looking st 40%. I wish to key in various parameters. I mean to say this exam is an important one in India. And most of the senior folks or coaching institutes are able to predict with their 10 years of experience. I have found that there is a pattern based on my research and connections with people in the coaching industry who coach students to write this exam. Just dont know how to program this, thats all. Hmm. Anyways will try out something.
@oudalally ok! Wont post. Thanks for the suggestion.
@gitreflog. Thanks for the reply. As i was writing i came up with this plan.
1. Retrieve all newspaper articles of one respected newspaper. And dump that day wise ordered monthwise to a db. Do this for 10 years.
2. Take the old 10 years previous questions. And do some ml stuff and find out how both correlates. For example, if i have year 2012s questons i will find out where in the db the topics were triggered and on what date. And making the ml model understand this character.
3. Like this i will have 10 trends in a ml model.
4. Finally i do reverse. This time i give it like the latest one year newspaper articles. And this model will predict based on past experience characters the topics that might be triggered. Also, You can increase accuracy by removing topics that have come already. So the amount needed to study reduces drastically.
So what do u think? Any ideas?
Wow thats is such sweet story indeed. I wish my sis understood this. After giving her a laptop with ubuntu she comes to after a week and says "I cant find the start button. Is this windows 8?". I can understand the start button part. But windows! Common sis. Its linux. Its on the top left side. Lols.