Χρήση ανατροφοδοτούμενου νευρωνικού δικτύου και χρονοσειρών για την πρόβλεψη δεδομένων
Using recurrent neural networks and time series for data prediction
View/
Keywords
Νευρωνικά δίκτυα ; Χρονοσειρές ; Δεδομένα twitter ; Πρόβλεψη ; LTSM ; Neural networks ; Time series ; Twitter data ; PredictionAbstract
There is no doubt that knowing the future can be the greatest advantage in life. A
person who has the ability to predict what the future holds can approach their goals
more effectively. There is also no doubt about the contribution of neural networks in
today's era and the numerous research applications they have provided. Neural
networks are powerful tools for prediction through time series analysis, and this
forms the core of this thesis.
Although we know that it is impossible to fully know the future, we can try to predict
what is coming. Both neural networks and social network analysis are distinct fields;
however, they can be correlated in various ways, as will be demonstrated in this
approach. The focus will be on analyzing and clarifying the structure of social
networks. Twitter, as one of the major pillars of social networks, will be the data
source for implementing this research design to achieve the goals of this work.
In this thesis, the following tasks will be carried out: visualization of the daily volume
of posts, limiting the set of users to those who have posted more than a certain
number of tweets, predicting the number of tweets published every minute based
on the data from the previous 24 hours, and finally, predicting the number of tweets
per minute based on historical data using the Long Short-Term Memory (LSTM) deep
learning model. The methods for time series prediction using a feedback neural
network will be disclosed, along with the process of creating the dataset and the
results of the two case studies.
Postgraduate Studies Programme
Προηγμένα Συστήματα Πληροφορικής - Ανάπτυξη Λογισμικού και Τεχνητής ΝοημοσύνηςDepartment
Σχολή Τεχνολογιών Πληροφορικής και Επικοινωνιών. Τμήμα ΠληροφορικήςNumber of pages
99Language
GreekCollections
Except where otherwise noted, this item's license is described as
Αναφορά Δημιουργού-Μη Εμπορική Χρήση-Όχι Παράγωγα Έργα 3.0 Ελλάδα
Αναφορά Δημιουργού-Μη Εμπορική Χρήση-Όχι Παράγωγα Έργα 3.0 Ελλάδα