Skip to content

Project1.md

Latest commit

 

History

History
171 lines (129 loc) · 4.05 KB

Project1.md

File metadata and controls

171 lines (129 loc) · 4.05 KB

Project 1: Diabetes Prediction

In this project our goal was to answer this question: "Can you build a machine learning model to accurately predict whether or not the patients in the dataset have diabetes or not?"

Dataset:

Link to dataset

Background Information

Variables Definition

  1. Pregnancies: Number of times patient was pregnant
  2. Glucose: Concentration of glucose after 2 hours [Used an oral glucose tolerance test]
  3. Blood Pressure: Diastolic Blood Pressure (mmHg)
  4. Skin Thickness: Skinfold thickness in triceps (mm)
  5. Insulin: 2-hour serum insulin (mmU/mL)
  6. BMI: Body Mass Index calculated using the formula (weight(kg)/ (height(m)^2))
  7. Diabetes Pedigree Function: Function that scores the likelihood of diabetes based on family history
  8. Age: How old they are Years
  9. Outcome: Boolean value 0 or 1

Data and Visuals

Single Variable Analysis

There are a total of 768 observations

Pregnancies:

Statistic Value
Mean 3.845
Standard Deviation 3.369
Minimum Value 0.000
Maximum Value 17.000
1st Quartile 1.000
2nd Quartile 3.000
3rd Quartile 6.000

Glucose:

Statistic Value
Mean 120.890
Standard Deviation 31.970
Minimum Value 0.000
Maximum Value 199.000
1st Quartile 99.000
2nd Quartile 117.000
3rd Quartile 140.250

Blood Pressure:

Statistic Value
Mean 69.105
Standard Deviation 19.350
Minimum Value 0.000
Maximum Value 122.000
1st Quartile 62.000
2nd Quartile 72.000
3rd Quartile 80.000

Skin Thickness:

Statistic Value
Mean 20.536
Standard Deviation 15.950
Minimum Value 0.000
Maximum Value 99.000
1st Quartile 0.000
2nd Quartile 23.000
3rd Quartile 32.000

Insulin:

Statistic Value
Mean 79.790
Standard Deviation 115.240
Minimum Value 0.000
Maximum Value 846.000
1st Quartile 0.000
2nd Quartile 30.500
3rd Quartile 127.250

BMI:

Statistic Value
Mean 31.990
Standard Deviation 7.884
Minimum Value 0.000
Maximum Value 67.100
1st Quartile 27.300
2nd Quartile 32.000
3rd Quartile 36.000

Diabetes Pedigree Function:

Statistic Value
Mean 0.471
Standard Deviation 0.331
Minimum Value 0.078
Maximum Value 2.420
1st Quartile 0.244
2nd Quartile 0.373
3rd Quartile 0.626

Age:

Statistic Value
Mean 33.240
Standard Deviation 11.760
Minimum Value 21.000
Maximum Value 81.000
1st Quartile 24.000
2nd Quartile 29.000
3rd Quartile 41.000

Outcome:

Statistic Value
Mean 0.348
Standard Deviation 0.476
Minimum Value 0.000
Maximum Value 1.000
1st Quartile 0.000
2nd Quartile 0.000
3rd Quartile 1.000

Bivariate Analysis:

Data Visualization

Results

Conclusion

Authors:

Team Advisor: Kyle Shannon

Team Leader and Lead Programmer: Seshu Mallina

Reserach Leader: Brandon Hui

Research Assistant and Editor: Marcus Cheung

Programming Assistants: Charvi Shukla and Garrett Lam

Acknowledgemnets

We would like to acknowledge our project advisor Professor Kyle Shannon of UCSD Cognitive Science deparment for his feedback and support throughout the course of the project.