Kriseledelse-bruk av sosiale medier i krisesituasjoner

Mehdi Ben Lazreg disputerer for ph.d.-graden med avhandlingen “A Neural Network-Based Situational Awareness Approach for Emergency Response” tirsdag 28. april 2020.

Han har fulgt doktorgradsprogrammet ved Fakultet for teknologi og realfag med spesialisering i IKT.

Slik oppsummerer kandidaten selv avhandlingen:

A Neural Network-Based Situational Awareness Approach for Emergency Response

In recent years, social media has become a part of daily life for many of us, and increasingly an open medium for emergency communications.

Emergency management services use social media to inform the public about the status of an emergency and what precaution the public needs to take. Also, the citizens use these online mediums to check on the safety of loved ones. Finally, eyewitness and affected individuals share their observations, concerns, and challenges they face during an emergency.

That information can potentially benefit emergency management services to improve situational awareness. However, emergency management services are still reluctant to use social media as a source of information.

Social media challenges

This thesis studies social media analysis platforms and the reason behind the emergency management services reluctance.

We identify two main challenges:

• The first challenge is the language used on social media, which includes misspellings, leetspeak, and abbreviations. This myriad of language usages requires a specific and adaptive normalization technique.
• The second challenge is that most of the dominant platforms try to find accurate ways of extracting as much information related to the crisis as possible. Consequently, they do not explicitly address the specific information-requirements of the time-constrained emergency personnel. This vast amount of supplied information causes an overflow of information, and in many aspects, renders the analytics platforms less useful.

Proposals to solve the challenges

As a mitigation to the first challenge, we propose a string metric that embraces similarities between text strings based on both the character similarities between the words and the context of these words.

For the second challenge, we propose an intelligent information retrieval framework for social media that, given the status of the emergency, provides the information most likely needed by the emergency management services.

The analytics framework we introduce combines two main components.

• The first component classifies social media messages into separate topics representing information required by emergency services during a specific situation.
• The second component decides which information to retrieve by learning what the emergency services need, based on the information available and the status of the emergency.

Slik gjør du som publikum:

Vi ber publikum om å ankomme digitalt tidligst ti minutter før oppgitt tid - det vil si  til prøveforelesningen 10:05 og disputasen tidligst kl 12:05. Etter disse klokkeslettene kan du når som helst forlate og komme inn igjen i møtet. Videre ber vi om at publikum slår av mikrofon og kamera, og har dette avslått under hele arrangementet. Det gjør du nederst til venstre i bildet når du er i Zoom. Vi anbefaler å velge «Speaker view». Dette velger du oppe til høyre i bildet når du er i Zoom.

Opponent ex auditorio: Disputasleder inviterer til spørsmål ex auditorio i innledningen i disputasen, med tidsfrister. Disputasleders e-post er tilgjengelig i chat-funksjonen under disputasen. Spørsmål om ex auditorio kan sendes til disputasleder Andreas Prinz.

Avhandlingen er tilgjengelig her:

 PhD Thesis - Mehdi Ben Lazreg