stm-feature-extraction

Paper title: Speed Transition Matrix Feature Extraction for Traffic State Estimation Using Machine Learning Algorithms
Extracting features from novel traffic data modeling technique called Speed Transition Matrix (STM). After the feature extraction, results are evaluated on different machine learning algorithms using labeled STMs.
You can learn more about STMs by reading these articles:

• Speed Transition Matrix: Novel road traffic data modeling technique
• Traffic State Estimation and Classification on Citywide Scale Using Speed Transition Matrices

Instalation

It is always recommended to use virtual environments and install packages from requirements.txt.
But, you can go ahead with python -m pip install numpy pandas matplotlib sklearn if you want.

Usage

1. Input data
   Data is not provided for this project but, you can contact me to send you some test data.
   Input data is composed of a list of dictionaries that contains STMs and metadata.
   This is the example of one entry in the dictionary:

{
  "stm": (ndarray),   # Two dimensional numpy array.
  "interval": (int),  # Time interval 0 to 7.
  "season": (str),    # If 'summer' data is collected in July and August if 'winter' then other months.
  "day": (str),       # 'working' or 'weekend'
  "origin_id": (int)  # Map-matched origin link id for STM
  "dest_id": (int)    # Map-matched destination link id for STM
}

2. Run
   Run python main.py.

Results

After running the main.py, two functions will run:

1. extract_features(data_path, save_path) will run the feature extraction step. All features will be explained in the article that will be published in Sep 2021.
2. get_comparison(save_path) will run the comparison of different ML algorithms. All results will be printed into a terminal.

How to cite

Text:
Tišljarić, L., Ribić, F., Majstorović, Ž., Carić, T. (2022). Speed Transition Matrix Feature Extraction for Traffic State Estimation Using Machine Learning Algorithms. In: Petrović, M., Novačko, L., Božić, D., Rožić, T. (eds) The Science and Development of Transport—ZIRP 2021. Springer, Cham. https://doi.org/10.1007/978-3-030-97528-9_5

.bib:
@Inbook{Tišljarić2022, author="Ti{\v{s}}ljari{'{c}}, Leo and Ribi{'{c}}, Filip and Majstorovi{'{c}}, {\v{Z}}eljko and Cari{'{c}}, Ton{\v{c}}i", editor="Petrovi{'{c}}, Marjana and Nova{\v{c}}ko, Luka and Bo{\v{z}}i{'{c}}, Diana and Ro{\v{z}}i{'{c}}, Tomislav", title="Speed Transition Matrix Feature Extraction for Traffic State Estimation Using Machine Learning Algorithms", bookTitle="The Science and Development of Transport---ZIRP 2021", year="2022", publisher="Springer International Publishing", address="Cham", pages="61--74", isbn="978-3-030-97528-9", doi="10.1007/978-3-030-97528-9_5", url="https://doi.org/10.1007/978-3-030-97528-9_5" }

Contact and connect

Leo Tisljaric