Fake-news
所属分类:仿真建模
开发工具:Jupyter Notebook
文件大小:42754KB
下载次数:0
上传日期:2020-07-21 08:57:44
上 传 者:
sh-1993
说明: 假新闻模拟:SLAPP中基于代理的转发网络模型
(Fake news simulation: agent-based model of retweet network in SLAPP)
文件列表:
Fake-news-Mas (0, 2020-07-21)
Fake-news-Mas\$$slapp$$ (0, 2020-07-21)
Fake-news-Mas\$$slapp$$\ActionGroup.py (338, 2020-07-21)
Fake-news-Mas\$$slapp$$\ModelSwarm.py (24474, 2020-07-21)
Fake-news-Mas\$$slapp$$\ObserverSwarm.py (5776, 2020-07-21)
Fake-news-Mas\$$slapp$$\Pen.py (458, 2020-07-21)
Fake-news-Mas\$$slapp$$\Tools.py (9821, 2020-07-21)
Fake-news-Mas\$$slapp$$\Tools_Addendum.py (1451, 2020-07-21)
Fake-news-Mas\$$slapp$$\__pycache__ (0, 2020-07-21)
Fake-news-Mas\$$slapp$$\__pycache__\ActionGroup.cpython-37.pyc (708, 2020-07-21)
Fake-news-Mas\$$slapp$$\__pycache__\ModelSwarm.cpython-37.pyc (10014, 2020-07-21)
Fake-news-Mas\$$slapp$$\__pycache__\ModelSwarm.cpython-38.pyc (10092, 2020-07-21)
Fake-news-Mas\$$slapp$$\__pycache__\ObserverSwarm.cpython-37.pyc (3346, 2020-07-21)
Fake-news-Mas\$$slapp$$\__pycache__\Pen.cpython-37.pyc (975, 2020-07-21)
Fake-news-Mas\$$slapp$$\__pycache__\Tools.cpython-37.pyc (5272, 2020-07-21)
Fake-news-Mas\$$slapp$$\__pycache__\Tools.cpython-38.pyc (5120, 2020-07-21)
Fake-news-Mas\$$slapp$$\__pycache__\agTools.cpython-37.pyc (1979, 2020-07-21)
Fake-news-Mas\$$slapp$$\__pycache__\graphicControl.cpython-37.pyc (2124, 2020-07-21)
Fake-news-Mas\$$slapp$$\__pycache__\txtxFunctions.cpython-37.pyc (1479, 2020-07-21)
Fake-news-Mas\$$slapp$$\agTools.py (1357, 2020-07-21)
Fake-news-Mas\$$slapp$$\convert_txtx_txt.py (2869, 2020-07-21)
Fake-news-Mas\$$slapp$$\convert_xls_txt.py (5400, 2020-07-21)
Fake-news-Mas\$$slapp$$\graphicControl.py (4181, 2020-07-21)
Fake-news-Mas\$$slapp$$\txtxFunctions.py (1888, 2020-07-21)
Fake-news-Mas\.ipynb_checkpoints (0, 2020-07-21)
Fake-news-Mas\.ipynb_checkpoints\iRunShellOnline-checkpoint.ipynb (902, 2020-07-21)
Fake-news-Mas\Analisi 1 (0, 2020-07-21)
Fake-news-Mas\Analisi 1\.ipynb_checkpoints (0, 2020-07-21)
Fake-news-Mas\Analisi 1\.ipynb_checkpoints\Power_Law_Analysis-checkpoint.ipynb (212287, 2020-07-21)
Fake-news-Mas\Analisi 1\Power_Law_Analysis.ipynb (1532825, 2020-07-21)
Fake-news-Mas\Analisi 1\information_network_finale.txt (5143393, 2020-07-21)
Fake-news-Mas\Analisi 2 (0, 2020-07-21)
Fake-news-Mas\Analisi 2\.ipynb_checkpoints (0, 2020-07-21)
Fake-news-Mas\Analisi 2\.ipynb_checkpoints\Lecture_Analysis-checkpoint.ipynb (15584, 2020-07-21)
Fake-news-Mas\Analisi 2\Lecture_Analysis.ipynb (42867, 2020-07-21)
Fake-news-Mas\Analisi 2\data.csv (3318367, 2020-07-21)
Fake-news-Mas\Analisi 2\data_init.csv (3217931, 2020-07-21)
Fake-news-Mas\Analisi 2\dati precedenti (0, 2020-07-21)
Fake-news-Mas\Analisi 2\dati precedenti\data_1.csv (311919, 2020-07-21)
... ...
# Fake-news
Stiamo realizzando un modello ad agenti in SLAPP, per simulare la diffusione di fake news su Twitter, nello scenario delle presidenziali
USA del 2016.
![alt text](https://dynaimage.cdn.cnn.com/cnn/q_auto,h_600/https%3A%2F%2Fcdn.cnn.com%2Fcnnnext%2Fdam%2Fassets%2F180126190057-08-white-house-mueller-denial-quotes-restricted.jpg)
## Il modello
I nodi rappresentano gli agenti, cioe gli account twitter.
Follower network: i link rappresentano il follow, cioe se A -> B, vuol dire che A e seguito da B.
Information network: I link rappresentano i retweet, cioe se A -> B, se l'account A ha retwettato un tweet dell'account B.
1. Si parte da un network random per quanto riguarda il follower network, scale free e diretto, con un numero fissato di nodi.
2. Si impostano le regole della dinamica.
3. Si valuta l'evoluzione dell' information e del follower network nel tempo.
4. Analisi dei dati.
## Dataset sul network di retweet
Per quanto riguarda i dati e le regole della dinamica, ci basiamo sul seguente [paper](https://www.nature.com/articles/s41467-018-07761-2) consultabile gratuitamente.
Vorremmo arrivare a riprodurre i seguenti [risultati](https://www.nature.com/articles/s41467-018-07761-2/tables/2)
## Sviluppo
### Fase 1: costruzione della rete ... completata
- Costruire i nodi ... done
- Implementare il preferential attachment ... done
- [Forest Fire](https://arxiv.org/pdf/physics/0603229.pdf) per il follower network.
- [Forest Fire Algorithm](http://snap.stanford.edu/snappy/doc/reference/GenForestFire.html)
- [Modello](https://ccl.northwestern.edu/netlogo/models/PreferentialAttachment) in Netlogo
### Fase 2: costruzione degli agenti
- Assegnare lo score agli agenti ... done
- Definire gli agenti nel file di testo ... done
- Implementare la verifica dello scale free: ... done
[Verificare che un network sia scale free](https://stackoverflow.com/questions/49908014/how-can-i-check-if-a-network-is-scale-free)
- Verificare che la rete sia scale free:
[Kolmogorov Smirnov test](https://en.wikipedia.org/wiki/Kolmogorov%E2%80%93Smirnov_test)
### Fase 3: meccanismi di interazione
- Implementare meccanismo di creazione della news ... done
- Implementare trasferimento della news ... done
- Implementare integrazione della news ... done
### Fase 4: regole della dinamica
- Associare ad ogni link un certo peso
[Network pesati](https://networkx.github.io/documentation/stable/auto_examples/drawing/plot_weighted_graph.html)
-[Regole della dinamica](https://docs.google.com/document/d/1kIeEAsEj68Kzrlez-EyenRuGMm5y2Hc0zdkFIFzppUE/edit?usp=sharing):
- Creazione bot ... done
- Meccanismi di interazione bot ... done
- Creare gli agenti (con i diversi score) rispettando le [proporzioni](https://docs.google.com/spreadsheets/d/12tik2m_w-y7LoZE5tSWOoTFDNzIT6hjsoO4Oaqabnoc/edit#gid=0) indicate in tabella
- Frequenza di creazione delle news rispettando le proporzioni ... done
### Fase 5: analisi dati
- Impostare l'analisi dati: conviene scrivere i dati in un file separato ... done
[Scrivere un file csv](https://www.programiz.com/python-programming/writing-csv-file)
[Scrivere un file da python](https://www.w3schools.com/python/python_file_write.asp)
- Network Analysis: ... done
[Power law fitting](https://github.com/micheletizzoni/Complexity-in-social-systems/blob/master/2-networkx/nb04_powerlaw_fitting.ipynb) //
[Clustering](https://github.com/micheletizzoni/Complexity-in-social-systems/blob/master/2-networkx/nb05_network_analysis.ipynb)
//[Barabasi power law fitting](https://barabasi.com/f/623.pdf) vedere pagina 44.
## Run it online with Binder
[Binder](https://mybinder.org/v2/gh/alessandrotofani/Fake-news/follower?filepath=iRunShellOnline.ipynb)
## Useful links
- [Regole della dinamica](https://docs.google.com/document/d/1kIeEAsEj68Kzrlez-EyenRuGMm5y2Hc0zdkFIFzppUE/edit?usp=sharing)
- [Proporzioni](https://docs.google.com/spreadsheets/d/12tik2m_w-y7LoZE5tSWOoTFDNzIT6hjsoO4Oaqabnoc/edit#gid=0)
- [md file cheatsheet](https://github.com/adam-p/markdown-here/wiki/Markdown-Cheatsheet)
## Un'altra immagine su Trump, why not?
![alt text](https://images.axios.com/YWMPTIF_Za_LAeU4cHUJwxe1R0M=/1920x1080/smart/2017/12/15/1513303959471.png)
近期下载者:
相关文件:
收藏者: