Estimating stock price correlations using Wikipedia
summary
Claims that Wikipedia text and links can estimate future stock correlations for DAX companies using concrete tools and numbers: article vectors via wiki2vec/word2vec, cosine similarity, graph distances and personalized PageRank, pricing from pandas datareader with adjusted close returns, and analysis in scikit-learn and networkX. Graph/category links prove too broad, vector text similarity yields modest signal with R^2 ≈ 0.2 after removing a Commerzbank outlier; code is on GitHub.
AI-written summary · about
chapters (12)
- 0:00Wikipedia and stock correlations
- 3:34Unconventional finance data
- 10:25Graph similarity from Wikipedia
- 13:16Pricing data and returns
- 17:01Initial results and findings
- 18:58Outliers in correlation analysis
- 21:08Wikipedia similarity effectiveness
- 23:16Finance applications
- 27:55Q&A proprietary algorithms
- 31:50Q&A Wikipedia timing
- 34:14Q&A modeling limits
- 35:36Q&A similarity measures
related talks
31mEstimating stock price correlations using Wikipedia
28mNetworks meet Finance in Python
35mCan Memes Predict Markets? Using Social Media Sentiment for Stock Forecasting
06mIgnite: Wild Correlations and more
30mIncreasing the trading prediction by mining aggregated human texting messages
59mExploring Wikipedia with Apache Spark: A Live Coding Demo
41mManaging Connected Big Data in Art with Neo4j Graph Database
24mHands-On Network Science