PASSENGER DESTINATION ESTIMATION USING PREDICTIVE MODEL WITH SMART CARD DATA
The Automated Fare Collection (AFC) system is Intelligent Transportation System which is popularly applied by public transport operators. In addition to facilitating the collection of tariffs, the data collected by this system is very useful in planning and public transportation strategies. But the...
Saved in:
| Main Author: | |
|---|---|
| Format: | Theses |
| Language: | Indonesia |
| Online Access: | https://digilib.itb.ac.id/gdl/view/36870 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Institution: | Institut Teknologi Bandung |
| Language: | Indonesia |
| Summary: | The Automated Fare Collection (AFC) system is Intelligent Transportation System which is popularly applied by public transport operators. In addition to facilitating the collection of tariffs, the data collected by this system is very useful in planning and public transportation strategies. But the AFC applied does not record all passenger transactions. This has made it difficult to know the needs and demands of public transportation.
In this study, propose a prediction model to estimate the purpose of passenger bus rapid transit (BRT) with smart card transaction data. Prediction models are built using decision tree and K-nearest neighbor (KNN) classification algorithms. The results of passenger destination predictions can be used to complete the missing transaction data in order to build an origin-destination matrix that can present the number of BRT passenger requests.
The data set used in this study is the smart card transaction data of BRT public transportation users with a span of one month. From the experiments conducted, it is known that individual travel information is the most influential thing on the predicted results of passenger destinations. The decision tree algorithm provides the results of the prediction of the destination stop better than the KNN algorithm, namely with a 52.2% f-measure value versus 49.3%. |
|---|