topic_immigration_it

This is a BERTopic model built as part of the European project VALAWAI and designed for predicting the topic distribution of immigration-related content in Italian language.
The model includes a pointer towards the model to be loaded in with SentenceTransformers, i.e., "sentence-transformers/paraphrase-multilingual-mpnet-base-v2".
The model was fine-tuned on a comprehensive set of tweets representing the information provided by both political entities and news sources on the immigration subject during the 5-year period 2018-2022.

Usage

To use this model, please install BERTopic:

pip install -U bertopic

You can use the model as follows:

from bertopic import BERTopic
topic_model = BERTopic.load("brema76/topic_immigration_it")

topic_model.get_topic_info()

example_text = "Questo è un esempio di testo sul topic immigrazione, subtopic sbarchi e accoglienza."
topics, probs = topic_model.transform(example_text)

probs = probs / probs.sum()

Topic overview

• Number of topics: 36
• Number of training documents: 159408

Click here for an overview of all topics.

Topic ID	Topic Keywords	Topic Frequency	Label
0	salvini - immigrato - ong - straniero - mare	77801	Crime
1	italia - sicilia - italiano - meloni - aquarius	27045	Reception
2	libia - libico - seawatch - tunisia - tunisino	18643	Sea rescue
3	ucraino - trump - profugo - polonia - russia	10779	Border crisis
4	sardegna - video - lampedusa - notte - algerino	3478	Landings
5	papa - chiesa - francesco - vescovo - papafrancesco	3395	The Church's view
6	coronavirus - positivo - virus - ospedale - contagio	2308	Coronavirus and its spread
7	hotspot - protestare - collasso - centro - lampedusa	2298	First reception centers
8	camion - arrestato - scafista - arresto - furgone	1624	Illegal immigration
9	incendio - baobab - tendopoli - moria - fiamma	1046	Shanty towns
10	reddito - pensione - cittadinanza - euro - investimento	1028	Economy
11	afghanistan - afgano - talebani - kabul - profugo	1011	Humanitarian corridors for refugees
12	turismo - turista - vacanza - estate - presenza	798	Tourism and vacations
13	incinta - bambino - donna - neonato - bimbo	776	Pregnancy, parenthood, and children
14	scuola - università - studente - tutore - lingua	651	Education and School-related themes
15	vaccino - vaccinato - vaccinale - vaccinare - covid	650	Vaccinations and EU digital Covid certificate
16	musumeci - ordinanza - islam - islamico - musulmano	621	Islam
17	alarmphone - egeo - naufragio - bambino - pericolo	552	Shipwrecks
18	stupro - sessuale - stuprato - violenza - stuprare	519	Sexual violence
19	multa - ong - decreto - amp - erostraniero	517	NGOs regulation
20	razzismo - razzista - odio - razziale - insulto	422	Racism and Hatred
21	agricoltura - schiavo - schiavitù - agricolo - schiavismo	409	Illegal work and exploitation
22	tubercoloso - malattia - tbc - salute - pandemia	361	Disease transmission
23	africa - africano - mission - continente - sviluppo	353	Cooperation and Development in Africa
24	asilo - giudice - richiedente - tribunale - ricorso	350	Right to asylum
25	dublino - regolamento - riforma - trattato - superare	310	EU regulation
26	fake - propaganda - fakenews - news - sapevatelo	285	Spread of fake news
27	droga - cocaina - hashish - eroina - marijuana	266	Drugs and Drug Dealing
28	film - miglior - oscar - sorrentino - dogman	263	Movies
29	gay - lgbt - omosessuale - cassazione - lgbti	149	LGBTQ+ and sexual minorities' rights
30	qatar - qatargate - panzeri - fifa - mondiale	129	Quatargate
31	perugia - università - suarez - rettore - universitàstranieri	127	Luis Suarez's Italian exam scam
32	brexit - regno - unito - inglese - qualificato	120	Brexit
33	basket - calcio - campionato - squadra - giocatore	117	Sports
34	profugo - volontario - gualzetti - sottoscrizione - accoglienza	107	Donations
35	matrimonio - combinato - finto - nozze - fittizio	100	Marriages and Fake Unions

Training hyperparameters

• calculate_probabilities: True
• language: None
• low_memory: False
• min_topic_size: 200
• n_gram_range: (1, 1)
• nr_topics: auto
• seed_topic_list: None
• top_n_words: 15
• verbose: True

Framework versions

• Numpy: 1.23.5
• HDBSCAN: 0.8.33
• UMAP: 0.5.3
• Pandas: 2.0.3
• Scikit-Learn: 1.3.0
• Sentence-transformers: 2.2.2
• Transformers: 4.33.1
• Numba: 0.56.4
• Plotly: 5.16.1
• Python: 3.10.11