Ranzcr

강의 홍보

• 취준생을 위한 강의를 제작하였습니다.
• 본 블로그를 통해서 강의를 수강하신 분은 게시글 제목과 링크를 수강하여 인프런 메시지를 통해 보내주시기를 바랍니다.
  • 스타벅스 아이스 아메리카노를 선물로 보내드리겠습니다.
• [비전공자 대환영] 제로베이스도 쉽게 입문하는 파이썬 데이터 분석 - 캐글입문기

Competition

• https://www.kaggle.com/c/ranzcr-clip-catheter-line-classification

Intro

• Thanks to RANZCR/resnext50_32x4d starter [training]
  • Please visit here and upvote

import os

import pandas as pd

from matplotlib import pyplot as plt
import seaborn as sns

Check File Size

• Check Each Size of Dataset Folder in this competition
  • train_records = 4.5GB
  • test_tfrecords = 0.5MB
  • train (image data) = 6.5GB
  • test (image data) = 0.8MB

import os

def get_folder_size(file_directory):
  # file_list = os.listdir(file_directory)
  dir_sizes = {}
  for r, d, f in os.walk(file_directory, False):
      size = sum(os.path.getsize(os.path.join(r,f)) for f in f+d)
      size += sum(dir_sizes[os.path.join(r,d)] for d in d)
      dir_sizes[r] = size
      print("{} is {} MB".format(r, round(size/2**20), 2))      
  
base_dir = '../input/ranzcr-clip-catheter-line-classification'
get_folder_size(base_dir)

../input/ranzcr-clip-catheter-line-classification/test is 805 MB
../input/ranzcr-clip-catheter-line-classification/test_tfrecords is 555 MB
../input/ranzcr-clip-catheter-line-classification/train_tfrecords is 4563 MB
../input/ranzcr-clip-catheter-line-classification/train is 6592 MB
../input/ranzcr-clip-catheter-line-classification is 12524 MB

Check train file

• Let’s descirbe train

train = pd.read_csv('../input/ranzcr-clip-catheter-line-classification/train.csv', index_col = 0)
test = pd.read_csv('../input/ranzcr-clip-catheter-line-classification/sample_submission.csv', index_col = 0)
display(train.head())
display(test.head())

Definitions of Variables

• What’s inside data?
  • StudyInstanceUID - unique ID for each image
  • ETT - Abnormal - endotracheal tube placement abnormal
  • ETT - Borderline - endotracheal tube placement borderline abnormal
  • ETT - Normal - endotracheal tube placement normal
  • NGT - Abnormal - nasogastric tube placement abnormal
  • NGT - Borderline - nasogastric tube placement borderline abnormal
  • NGT - Incompletely Imaged - nasogastric tube placement inconclusive due to imaging
  • NGT - Normal - nasogastric tube placement borderline normal
  • CVC - Abnormal - central venous catheter placement abnormal
  • CVC - Borderline - central venous catheter placement borderline abnormal
  • CVC - Normal - central venous catheter placement normal
  • Swan Ganz Catheter Present(??)
  • PatientID - unique ID for each patient in the dataset