Project: "Novelty Detection Framework"
Owner: "Giuseppe Pedone, Umberto Albertin"
Date: "2023:01"
All the code and models inside this repository are released under the MIT Licence.
MIT License
Copyright (c) 2023 Giuseppe Pedone, Umberto Albertin
Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions:
The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software.
THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
This NDF project is a way to detect nolveties/changes on new datasets, given a model pretrained on 'normal' datasets. It is composed by two parts: AI Unit and Maintenance Control Unit.
In order to install the framework, you need to perform the following tasks: - Install MongoDB on your system. - Launch the mongod deamon as superuser. - Install, using 'pip install -r requirements.txt'. - Install the last version of Boost Library on your system. - Install libmongocxx version 3.6.7 or later. - Install libbsoncxx version 3.6.7 or later. - Install libmongoc-1.0. - Install libbson-1.0. - Install the last version of OpenSSL.
The NDF configuration is based on local/remote json files, like configuration.json file found into the directory tree of of the application.
{
predictor_recover_directory:"./Recovery", //setting the path of Recovery directory
column_names:predicted_features, //setting the predicted features names
nmin_datasets_for_train:5, //set the minimum number of dataset to read before start the training process
nmin_dataset_for_test:2, //set the minimum number of dataset to read before start the testing process
nmax_dataset_for_test:2, //set the maximum number of dataset to read before start the testing process
mongo_string:mongodb://localhost:27017 //set the mongodb URI string used to connect to mongod daemon
raw_db:RawDB, //set the name of raw data database
info_db:InfoDB, //set the name of info data database
rawdataset_collection:Dataset, //set the name of raw data collection
trainset_collection:Trainset,, //set the name of trained data collection
testresult_collection:TestResult, //set the name of test result collection
performance_collection:Performance,, //set the name of performance collection
configuration_collection:Configuration, //set the name of configuration collection
windows:20, , //set the number of time consecutive samples considered into a record
random_state:0, , //set the initial random state of the Random Forest Regressor
n_estimators:6, //set the number of estimators of the Random Forest Regressor
max_features:2, //set the max features
trained:false, //identify the trained phase for AIU
tested:false, //identify the testing phase of AIU
dataloge:true, //set log enable
timestamp:2022-11-19T03:06:35.397+00:00, //set the time of current configuration
agents: //Identify the agents and their configurations. Set an Agent ID in in order to add a new agent
AG0:{
num_samples_read:10, //set the number of samples to read before performing an action
predictor:7, //set the predictor number that is managed by the agent, used to get properly data.
predicted_error_type:MAE, //set the type of error used by the agent for taking decisions, used to get properly data.
predicted_error_value:100.1, //set the error value used by the agent for time predictions, used to predict a time of error.
min_operative_threshold_error:20.2, //set the minimum error value to reach in order to activate the agent computation, if it is not reached, the agents do nothing.
max_operative_threshold_error:80.2, //set the maximum error value to reach in order to activate the agent computation, if it is overcomed, the agents do nothing.
min_num_of_regr_samples:10, //set the minimum number of samples to use for regression, if it is not reached, the prediction is not performed.
prevention_threshold_time:68000000000, //set the duration, in milliseconds, of enduser notification threshold. It sets the minimum time duration within a certain model changes have to occur in order to notify the enduser.
test_result_collection:TestResult, //set the name of test result collection
prediction_result_collection:Prediction, //set the name of prediction collection
database_name:InfoDB, //set the name of info data database
mongo_driver_remote_connection_type:mongodb, //set the mongodb connection type
mongodriver_remote_connection_host:localhost, //set the mongodb connection host
mongo_driver_remote_connection_port:27017, //set the mongodb connection port
agent_id:AG0, //set the agent id used to attach the right configuration to the right agent
agent_name:Agent 0, //set the agent name
step_run_time:5000, //set the agent execution step time. Agent will be executed and ready every t+Step milliseconds
priority:4, //set the execution order priority used to sort the executing agent queue.
stopped:false} //set the agent stop state. If it will be true, you need to activate the agent via code with agent.fActivate() function, called inside the Agent Manager costructor.
}
After setting the parameters into the local configuration.json file, or remotly on mongodb configuration collection, open a root terminal and run the following command:
cd <PATH TO ROOT OF NDF>
and the On Linux:
./Demo.sh
On Windows:
.\\Demo.bat
After the NDF is running, you can interact with NDFs AIU using keyboard commands. Type:
- 'R' on the AIU Shell to retrain the NDF.
- 'U' on the AIU Shell to update the NDF.
- 'S' on the AIU Shell to save a new NDF Configuration.
- 'T' on the AIU Shell to load the last NDF Configuration.
- 'P' on the Plot Shell to plot all NDF predictions.
- 'Q' on the Plot Shell to close the Plot Shell.
Restart the Demo if a Configuration file variation is performed.
Giuseppe Pedone, Umberto Albertin