Data Driven Diagnosis of Alzheimer's Disease
summary
A citizen science pipeline scrapes the Alzheimer’s Society forums with BeautifulSoup and produces user- and post-level dataframes (about 3,600 threads and 78,000 posts, ~30% started by sufferers). Text features include part-of-speech rates, cosine/edit-distance repeats, sentiment via textblob, readability via readability-lxml, and LSH-style comparisons; models use scikit‑learn random forests and SVM RBF kernels. User aggregation yields ~63% accuracy, post classification reaches 68% accuracy, and a prototype random forest–based early‑warning concept is produced.
AI-written summary · about
chapters (8)
- 0:00Alzheimer's diagnosis project intro
- 2:38Global impact of Alzheimer's
- 9:30Emails to forum data switch
- 16:48Labeling users and biases
- 22:06Early detection features
- 27:04Model performance and research
- 31:05Medical data science cautions
- 36:16Q&A
related talks
27mMachine Learning Methods in Detecting Alzheimer’s Disease from Speech and Language
38mEthical Artificial Intelligence to Advance Welcoming Communities and Identify Potential Communica...
32mAnalyzing social media for suicide risk using natural language processing
32mTriaging Feedback Form Data
01h 24mIntroductory tutorial on data exploration and statistical models
37mData Science for Digital Humanities: Extracting meaning from Images and Text
03h 51mMachine Learning Summit
39mSentiment Analysis, its techniques and applications