研究目的
Investigating the use of open-source software and machine learning techniques for the detection of Diabetic Retinopathy (DR).
研究成果
The use of open-source technology and machine learning techniques like BDT and DNN provides an effective and convenient method for the diagnosis of Diabetic Retinopathy. Despite the challenges with low statistics, the results are reasonably good.
研究不足
The study faced challenges with low statistics, as indicated by the overtraining plots of BDT and DNN, which affected the accuracy of the results.
1：Experimental Design and Method Selection:
Utilized open-source software for coding and applied machine learning algorithms like boosted decision tree and neural network on data from the UCI machine learning repository.
2：Sample Selection and Data Sources:
Used retinal images of diabetic retinopathy patients from the UCI repository.
3：List of Experimental Equipment and Materials:
Open-source software tools for machine learning.
4：Experimental Procedures and Operational Workflow:
Extracted features from retinal images using image processing algorithms and applied machine learning techniques for classification.
5：Data Analysis Methods:
Analyzed the performance of machine learning algorithms using ROC curves and overtraining checks.
独家科研数据包，助您复现前沿成果，加速创新突破
获取完整内容
