This repository contains code related to the following
- Neural Network Training and Testing (for Traffic Flow Prediction)
- Evaluating Performance of Other Regression Methods
- Traffic Flow Data Collected from PeMS
- Producing and Consuming Records to/from Kafka
- Experimental Results
All code related to the LSTM network is in the TrafficPrediction
folder.
Performed by evaluate_regression.py
in the TrafficPrediction
folder.
Contained in the data
folder
Relevant code is located in kafka_scripts
. Also contains information on how to run the experiment.
All latencies and corresponding analysis are located in the results
folder
Please visit each of these folders and see their README for more information. We will provide brief overviews for the folders not listed above. Note that the following folders are either no longer useful to this experiment, or only contain helper code.
Stores images that are helpful visualiziations of the neural net/system architecture
This folder is deprecated. A previous version of the experiments used NSQ instead of Kafka, and these escripts contained the NSQ equivalents of what is located in the kafka_scripts
folder. They may be used as reference, but should be updated should you wish to use them in an experimental setting
Contains data pertaining to detector locations that were not used in the later versions of the experiment
Contains LaTeX for the written report
Helper functions to generate periodic workloads
Random collection of utility functions