Comments on: Data Science Practice – Classifying Heart Disease

By: Kyle Hooks

Kyle Hooks — Mon, 08 May 2017 17:56:33 +0000

I found if you prune some less useful features you can get up to 78%. I haven’t tried testing various initialization values for a classifier yet, but viewing it as regression got me to a stdev of 0.22 after testing many parameter initialization values. I’m not sure how to test the “accuracy” here without some threshold value, but I’d say it’s 90% accurate by eyeballing it (if you round to nearest integer values).

By: dibya

dibya — Tue, 11 Apr 2017 06:19:00 +0000

hello very nice project for beginners. can you guide me how to proceed further? how to improve the accuracy??

By: AND

AND — Sat, 29 Oct 2016 19:40:40 +0000

heart_test_results[‘correct’] = heart_test_results[‘predict’] == goal_test_df[‘actual’]
File “/home/admin-pc/anaconda3/lib/python3.5/site-packages/pandas/core/ops.py”, line 735, in wrapper
raise ValueError(‘Series lengths must match to compare’)

I am unable to figure this error :/

By: Renee

Renee — Mon, 22 Aug 2016 05:31:38 +0000

In reply to aziz. It's not my dataset. If you're having trouble accessing it via the UCI respository, you should probably let them know. Sorry, but I don't have time to re-download and email datasets to individuals when they're available online.

By: aziz

aziz — Sun, 21 Aug 2016 04:53:22 +0000

Hello,
I couldn’t download the datase. can you mail me? [admin removed email]

By: Renee

Renee — Mon, 08 Aug 2016 14:24:30 +0000

In reply to Manasi.

Hm, I used the one that is here called processed.Cleveland.data, I think
https://archive.ics.uci.edu/ml/machine-learning-databases/heart-disease/

By: Manasi

Manasi — Mon, 01 Aug 2016 09:13:49 +0000

Hello,
I couldn’t download the dataset. I couldnt find the one with 76 attributes. Could you please pass over that particular dataset?

Manasi

By: Renee

Renee — Mon, 16 Mar 2015 04:07:28 +0000

In reply to Felipe Bormann. Hi Felipe, Great, I'm glad to have given you courage to post about your data science adventures! Post a link here when you get that started! You can also follow me on twitter at @becomingdatasci. Renee

By: Felipe Bormann

Felipe Bormann — Thu, 12 Mar 2015 22:18:32 +0000

Congratulations, I’m starting to learn data science as well, like, I’ve just started computer science but I have a feeling it is the right field for me, I’ll start posting my adventures really soon, you have gave me the courage, please, keep up the good work, is there somewhere else I can follow you?. Have a great day.

Oh, I’m from Recife, Brazil o/!

By: Renee

Renee — Mon, 23 Feb 2015 02:44:24 +0000

In reply to Natalie. Hi, thanks Natalie! I haven't used dplyr (my R experience is very limited), and I was comparing to the "manual" (without a package to help) data importing I had done in past projects when I first learned Python, where I had to manipulate the columns using indexes to get what I wanted into the final set, and where the importing was less straightforward. You can see some of that here: https://www.becomingadatascientist.com/2014/04/10/ml-project-2-post-2/ I enjoyed being able to work with the columns by referencing their names. I know I just touched on the surface of what Pandas can do, and yes I am planning to do more practice projects like this soon!

By: Natalie

Natalie — Wed, 18 Feb 2015 08:19:39 +0000

Very interesting! Did you feel pandas was better with import and manipulation? Compared with, say, using the dplyr package for R? Or if you don’t use that, what are you comparing with? If you’ve got more exploratory posts like this in the pipeline I hope you post them soon, because I enjoyed reading this :)