| Name: | Description: | Size: | Format: | |
|---|---|---|---|---|
| 5 MB | Adobe PDF |
Advisor(s)
Abstract(s)
A construção de cidades inteligentes e sustentáveis engloba diferentes sectores, desde a
mobilidade, a energia ou qualquer serviço necessário à vida das pessoas.
Nesta dissertação o foco é a mobilidade dentro das cidades inteligentes, ou futuras
cidades inteligentes, e objetivo principal é desenvolver uma metodologia que possa fazer
a monitorização e gestão de transportes, principalmente de transportes públicos.
Utilizou-se dois algoritmos de machine learning neste projeto: K-means Clustering e
Support Vector Machine. O objetivo do primeiro algoritmo é identificar padrões de rotas
de transportes públicos. E o segundo serve para fazer a previsão do tempo de chegada do
transporte duma determinada rota.
O algoritmo de K-means consegue identificador o fluxo de passageiros e fazer a separação
de clusters com mais passageiros.
O algoritmo de SVM consegue fazer a previsão da hora de chegada do transporte com
um erro máximo de 2 minutos.
The construction of smart and sustainable cities encompasses different sectors, from mobility, energy or any service necessary for people’s lives. In this thesis the focus is on mobility within smart cities or future smart cities and the main objective is to develop an algorithm that can monitor and manage transport, especially public transport. Two machine learning algorithms were used in this project: K-means Clustering and Support Vector Machine. The objective of the first algorithm is to identify public transport routes behaviors and patterns. The second is used to predict the arrival time of transport on a given route. The K-means algorithm can identify the flow of passengers, and separate clusters with more passengers. The SVM algorithm is able to forecast the arrival time of the transport with a maximum error of 2 minutes.
The construction of smart and sustainable cities encompasses different sectors, from mobility, energy or any service necessary for people’s lives. In this thesis the focus is on mobility within smart cities or future smart cities and the main objective is to develop an algorithm that can monitor and manage transport, especially public transport. Two machine learning algorithms were used in this project: K-means Clustering and Support Vector Machine. The objective of the first algorithm is to identify public transport routes behaviors and patterns. The second is used to predict the arrival time of transport on a given route. The K-means algorithm can identify the flow of passengers, and separate clusters with more passengers. The SVM algorithm is able to forecast the arrival time of the transport with a maximum error of 2 minutes.
Description
Keywords
Monitorização de transportes Gestão de transportes Algoritmo de previsão Machine learning Support vector machine K-means clustering