This project is protected by a MIT License.
The objective of this project is to provide a mail analysis tool allowing to no longer rely the analysis (and therefore the responsibility) of the analysis of mail on the knowledge of prevention of the reader but also on a tool.
This tool is intended to be as discreet as possible, it will then perform a preprocessing of the mail before it is opened by the user. This treatment will then assign a note to the mail, a note that corresponds to the credibility of the mail and therefore to the confidence that we can have in him.
One of the tools we use in this project is a Gmail mailbox that serves as the entry point for emails and serves as a witness to the proper functioning of our methods.
This tool use ML (Machine Learning) and NN (Neural Network) technologies to analysis email. To run this project you need to install packages which are :
- numpy
pip install numpy - pandas
pip install pandas - sklearn
pip install sklearn - nltk
pip install nltk - ip2geotools
pip install ip2geotools - requests
pip install request - tensorflow or tensorflow_gpu (if you have a gpu compatible)
pip install tensorflowpip install tensorflow-gpuIf you decide to use your gpu please install Cuda (nowadays Cuda 10.0 is stabler than Cuda 10.*) - keras
pip install keras
Team is composed by 4 members which are :
- Alan Adotevi
- Quentin Boëns
- Antoine Genonceau
- Pierre Marez
All of them are students at ISEN Lille in 2019-2020