Some people say that in the 21st century, our lives are becoming more and more convenient, electronic communications are more and more developed, and there are more and more food. This is not fake at all. But the modern lifestyle also has its downsides. Work is getting busy, physical exercise is getting less and less, and body weight is increasing day by day. There is a disease called diabetes, have you heard of it? The number of diabetic patients in China has been increasing year by year, showing a significant upward trend. In recent years, diabetes drugs have become a big cake for pharmaceutical companies. Diabetes-related foods are also very expensive, and many patients who do not understand medical knowledge pay a huge price.
Some people say that in the 21st century, our lives are becoming more and more convenient, electronic communications are more and more developed, and there are more and more food. This is not fake at all. But the modern lifestyle also has its downsides. Work is getting busy, physical exercise is getting less and less, and body weight is increasing day by day. There is a disease called diabetes, have you heard of it?
In order to increase sales, the catering industry will increase the amount of food salt. This can satisfy customers' heavy tastes. 
In order to increase sales, beverage manufacturers add a lot of sugar (carbohydrates) to their beverages. Sugar can stimulate the brain and form positive feedback. The more you drink, the more you want to drink, forming a tolerance for sugar. 
996 is no stranger to programmers. He is too busy at get off work a day, so Ge You lie down after work, holding a mobile phone or watching TV. The weight is increasing day by day, what I want to say is diabetes is far from you? 
The truth is that China's diabetes data is shocking! The figure below shows that the estimated prevalence of diabetes in China in 2019 ranks second in the world. We are not the first in the world, are you happy? 
Unfortunately, China's population base is more than four times that of the United States, so the number of diabetes patients in China ranks first in the world. China is the largest drug development market for diabetes. More and more young people have joined the diabetes market and become cash cows for pharmaceutical companies. 
The number of diabetic patients in China has been increasing year by year, showing a significant upward trend. In recent years, diabetes drugs have become a big cake for pharmaceutical companies. Diabetes-related foods are also very expensive, and many patients who do not understand medical knowledge pay a huge price. 
Diabetes is a disease of wealth. It cannot be cured completely after suffering from it and can only be controlled with daily medication. And there are many complications related to diabetes. Every 8 seconds, one person dies of diabetes and its complications. If left untreated, diabetes can cause many complications. Acute complications include diabetic ketoacidemia and high blood pressure and high blood sugar non-ketoacid coma; serious long-term complications include cardiovascular disease, stroke, chronic kidney disease, diabetic foot, and retinopathy. 
Doctors recommend more for diabetics, eat less sugar-rich foods, exercise more, and rest more. . . . But which recommendation is particularly important, can it be quantitatively analyzed? The answer is yes, you give me data, and I give you the answer.
Artificial intelligence machine learning can help medical researchers discover more diabetes-causing factors and build models to predict patients' blood sugar. The blogger uses python to build a diabetes blood glucose index prediction model. According to the age, gender, blood pressure, BMI and other indicators provided by the user, it can predict whether you have diabetes. The modeling data comes from real diabetes clinical data in the United States, totaling 442 items. For the model, the amount of 442 pieces of data is somewhat small. If there are more than 1,000 pieces of data, it would be nice. 
The American team (Bradley Efron, Trevor Hastie, Iain Johnstone and Robert Tibshirani) researching this project has average model performance, MAE is about 41.9, r2 is 0.477. The blogger’s model has a MAE of 13.82 and r2 of 0.9388, which is much higher than that of the US team. The performance of the model is very good, and the prediction data error is small and more accurate. 
The blogger uses python language to establish a screenshot of part of the code of the diabetes prediction model, with a small amount of code, high efficiency, rapid modeling and quantitative analysis of pathogenic factors. 
After the program analyzes the correlation of variables, it automatically saves the results to excel for easy access to data in the future. We found that the correlation between S1 and S2 serum indicators is very high. The model only uses s1 or s2 variables, and the performance decline will not be too great. 
The good news is that most diabetes falls into the second category, which is preventable and controllable. As long as we are familiar with the pathogenic factors and establish good living habits, we can significantly reduce the probability of diabetes. This course analyzes the pathogenic factors of diabetes one by one and ranks them quantitatively. It is a very valuable course. 
Machine learning is a fascinating subject that allows us to be like Gandalf magicians and can predict the future. I hope my course can help diabetic patients, related research and development institutions, or students who are writing papers on this subject. I hope you will share this course to your circle of friends so that everyone can pay attention to diabetes prevention and control, reduce government medical budget expenditures, and benefit more people.
Welcome all students to study
Python machine learning-diabetes data mining :
https://edu.51cto.com/sd/5c2fb
The author introduces
Toby, a licensed financial company as a model verification expert, and the head of the data mining department of the largest medical data center in China! Cooperate with Chongqing Children's Hospital, Professor of Chinese Academy of Sciences, Cyberlan to maintain the chronic disease data mining project! Managed foreign pharmacopoeia databases such as Europe, America, Japan, China, India and Brazil, Martindale database, FDA solubility database, clinical trial database, WHO drug warning and other databases.
Course overview
Python machine learning actual diabetes data mining, using multiple regression algorithms to discover which factor is the most important pathogenic factor among age, gender, body mass index BMI, blood pressure, and six serum indicators. How are these variables related. The comprehensive performance of the curriculum building model is significantly higher than other Internet courses.
Applicable population:
Graduate students, doctoral dissertations, NCBI/SCI/Nature dissertations, python lovers, machine learning, bioinformatics, diabetes medical research institutions
The curriculum features
civilian prices, not purely commercial prices, so that poor students can learn the most advanced popular knowledge abroad. There is no need to spend tens of thousands to report sky-high prices for study classes, self-study can also grow.
Study plan and method
1. Guarantee 1-2 hours of study time every day, and it is estimated that you can study a complete course in 7-14 days.
2. The code practice of each lesson should be guaranteed. It is recommended not to copy and paste the code directly. It is very important for the brain to remember the code and consolidate knowledge.
3. In the second study, you should summarize the content of the previous lesson, and make notes if necessary to deepen your brain understanding.
4. If you don't understand the questions, please list them, and check them online first. If you can't find them, you can consult the teacher.
Course Catalogue
Hours 1 Introduction to all my courses
Class hours 2 Lecturer introduction-Twenty medical database leaders
Class hours 3 Diabetes classification_features_Prevention overview
Class hours 4 Machine learning model prediction of blood glucose indicators for diabetic patients
Chapter 2 Python programming environment construction
hours 5 Anaconda Quick Start Guide
Class hours 6 Anaconda download and installation
Class hours 7 Python third-party package installation (pip and conda install)
Chapter 3 Diabetes data mining
Class hours 8 Establish a linear regression model for diabetes prediction (linear regression)
Class hours 9 Download diabetes data methods (raw data and cleaned data)
Class hours 10 Linear Regression and Error
Class 11 Model Verification: Mean Square Error and Median Absolute Error
Class 12 multi-algorithm comparison, the model performance is increased by 2 times
Class 13 Raw data and processed data modeling performance comparison
Class 14 Diabetes pathogenic factors quantitative analysis_ gender, age , Blood pressure, BMI index
class hour 15 variable correlation analysis-the original s1 and s2 serum test showed high correlation.
Class hour 16 longevity-this course is your lifetime wealth,
class hour 17 BMI index python automatic calculation script
chapter 4 appendix
class hour 18 diabetes Chinese and English words Summary
class 19 Are diabetic patients more likely to contract the new coronavirus?
Lesson 20 LeastAngleRegression LeastAngleRegression
Python machine learning bioinformatics, blogger recording, 2k ultra-clear
https://edu.51cto.com/sd/3a516