- OUR Solution: a random forest classifier based on multiple features fusion
- First and foremost, thank my excellent teammates @solaris, @Mikan1206. Solaris's work for extracting the features from the text (nlp) and images(CV) impress me greatly. And no words can be found to describe the spirit of teamwork throught the four days. What binds us together is the the passsion for cracking the data and the love for writing codes. Still, there are regrets that we do not display the process and results in a more mathematical way.
- a random forest classifier based on multiple features fusion
- basically, this is the work combining NLP and CV for better understanding the spread of a alien species. Features are extracted in four different parts, text, images, comments and date.
- the complete python code for the five Tasks
- the dataset for completing the task
- Detectron2.ipynb do a image preprocessing job
- task1.ipynb complete a time series analysis for task 1
- Ramdomforest.ipynb
- Note that all the code were written in colab originally
- install jupter notebook or directly upload to colab
- git clone this repository
- unzip SIFT.zip
- task1.ipynb and Ramdomforst.ipynb is the main code file