#statstab #474 Linear Models with Heterogeneous Coefficients
Thoughts: Sometimes you need more complicated models even if identification gets messy.
#heterogeneity #modelling #nonlinear #economics #econometrics
https://vladislav-morozov.github.io/econometrics-heterogeneity/linear/linear-introduction.html
Job Alert
Three doctoral students (75%) for the ifo Center for International Economics
Deadline: 2026-01-06
Location: Germany, München
#Econometrics #Economics #EmpiricalEconomics #InternationalEconomics #PhD #hiring
🎉 Gretl 2025c is here!
Exciting updates to your favorite econometrics toolkit! Version 2025c brings powerful new features and improvements:
✨ New Features:
Gibbs sampler command for Bayesian analysis is available now!
🚀 Performance & Quality:
Faster forward stepwise regression
🎨 GUI Enhancements:
Better dbnomics search integration
Improved dark theme support
Full changelog:
https://gretl.sourceforge.net/ChangeLog.html
#gretl #econometrics #opensource #statistics #datascience #economics #timeseries
Recent @DSLC club meetings:
Advanced R: Function operators https://youtu.be/fTsquqgEeZw #RStats
From the @DSLC chives:
R for Data Science: Functions https://youtu.be/GyfC1MgP_fw #RStats
Econometrics with R: Linear Regression with One Regressor https://youtu.be/WiZEdxfJ6sY #RStats #econometrics
Visit https://dslc.video for hours of new #DataScience videos every week!
AI’s $1 trillion bet - is it an #AI bubble or dot-com bust? Global data‑center capital expenditure to power AI is projected to rise from roughly $430 billion this year to over $1.1 trillion by 2029 (which is equal to the GDP of the Netherlands). Why it matters:
We’re witnessing an infrastructure boom that echoes a familiar pattern in tech history but the question is whether it’s building toward lasting transformation or racing toward collapse.
Capital is pouring into data centers, cooling systems, power infrastructure, and networks at a staggering pace. Yet most AI applications haven’t proven they can generate sustainable revenue at scale.
BTW, this article is a good one, even though it bears the hallmarks of some AI input. I’m tired of hearing about #AISlop from folks who don’t even read materials to determine if they deliver meaningful content. I use AI as my research & brainstorming asst. It works.
https://www.jeffbullas.com/ai-bubble-or-dot-com-bust/ #ArtificialIntelligence #economy #econometrics #markets #finance #technology
AI’s growth is being built on concrete, chips, and power lines, not just code. But when $1 trillion flows into infrastructure before profits arrive, you have to ask: are we building the future, or another bubble?
👽 There it is! 👾 Our new and shiny paper for the Journal of Econometrics! 🤖
https://doi.org/10.1016/j.jeconom.2025.106107 🚀
In this paper:
✅ we provide general conditions for partial identification of Structural VARs through heteroskedasticity
✅ we show that it's great for analysing fiscal policy effects on the economy
✅ it's the methodological paper for my bsvars package
👇
Date: Wednesday, 1 October 2025, at 2:00 pm Venue: Seminar Room 237, DEM + Online via Teams Speaker: Claudia Pigini…Leggi tutto...
CPC-CG members Professor Jackie Wahba OBE and Professor Athina Vlachantoni have been announced as #REF 2029 Sub-panel members for #Economics and #Econometrics, and #SocialWork and #SocialPolicy, respectively.
They join CPC-CG Director Professor Jane Falkingham CBE who is Chair of Main Panel C– #SocialSciences. Full story: https://www.cpc.ac.uk/news/latest_news/?action=story&id=847
#researchexcellenceframework #demography #research #socialscience #ageing #migration #economist
從檢定發現美國失業率是廣義極值分配特性。同時，Gumbel,type I有機率密度函數，真實告訴你美國失業率的機率模型，而不是出一張圖代表存在機率模型。
以上這些方法都是超越傳統AI的數據分析方法，真正從數據本質出發打造精確統計模型，解決通用模型無法捕捉真實數據規律的難題，通過自動化建模過程揭示隱藏的數學規律。
你學的是落在哪種層次呢？
直線建模能做到，當然非線性的人工智慧自動化建模同樣能做到。數據規律的數學化、自動化(更新+建模+模擬)、強大而直觀的統計分析工具集成，統計學習達成。
其中一種非線性建模：https://x.com/meiyulee357/status/1963281302423241168
@academicchatter @econometrics @ida
#AI #數據分析 #失業 #美國 #經濟 #計量經濟 #modelling #econometrics #unemployment #Statistics #USA #dataanalysis
如何建構美國失業率的機率模型？機率模型最後要有數學式顯示，不能只是圖形。
1) 直方圖？別想了，沒有數學式。只是圖像視覺化，不是數據分析，也不是人工智慧該有的數據模型。【不合格】
2) 用直方圖的組中點和對應機率值？11組可以使用AI-based piecewise linear regression method的結果是兩段直線。整體的R2達73%。【合格】
3) 建立更多分組的直方圖產生組中點與機率值。運用AI-based piecewise linear regressin method，產生9段直線。整體的R2達93%。【合格】
4) 運用適合度檢定，檢定45種機率分配？發現美國失業率的機率模型服從Gumbel,type I(a=0.68,b=27.82)。根據a值升序模擬產生條件機率分配。【合格】
@academicchatter @econometrics @ida
#AI #數據分析 #失業 #美國 #經濟 #計量經濟 #modelling #econometrics #Statistics #artificialintelligence
Dr Mircea Zloteanu ❄️☃️🎄
Academic Europe
Gretl
DSLC Videos
Charlie McHenry
Tomasz Woźniak
Pietro Battiston
Centre for Population Change
AI數字分析領航者 
AI底層數據建模