The code from data collection to the model building is present in the ipynb notebook in the Python code folder

• We imported the necessary libraries
• Mount to the google drive
• Change the path to the folder where the dataset is downlaoded in the drive
• Load the dataset in the mat format and convert it into csv format and load the dataframe
• In the dataframe the column names are actually the indexes and the indexes are the columns. We will transpose the DataFrame to make them proper.
• Rename the columns in the dataframe
• Synthetic data generation
• Exploratory data analytics(Histograms,Correlation analysis,Heatmaps)
• Outlier Detection and removal
• Model building for predicting final positions and velocities
• Model building for detecting anomaly along with addressing class imbalance problem
• Select the best model in terms of accuracy and F1 score( Decision Tree Classifier)
• Link to previous deployment on streamlit

• Dockerfile is the final Docker image
• app.py is the Flask application built
• templates folder and static/css folders contains the html and css files related to the Flask application
• clf_model is the DecisiontTree classification model
• model.h5 and model.json are the model weights of ANN model built for predicting final positions and velocities in h5 and json format respectively
• requirements.txt file is the libraries that will be installed during docker deployment

• First, using the Flask a simple Web Application was made.
• Download a ubuntu virtual machine and install docker using official documentation
• Setup Docker Daemon
• Go to project folder where Dockerfile is present. Dockerfile is used for building of the docker image.
• In my case it it btp_project_deployment
• Use the below command for build the docker image

docker build -t <name> .

• Use the following command for viewing all the docker images

docker images

• Run the docker container using the below command

docker run -d -p 5000:5000 <name>

• Open the flask app running on port 5000 in any web browser available

localhost:5000

• Give the inputs and click on predict and the app will run the model and predict if there is any anomaly or not
• 1 indicate anomaly and 0 indicate no anomaly
• We can know which docker containers are running using

docker ps

• Stop the docker container from running by using

docker stop <id>

• Delete the image by using the following command

docker image rm -f <image name>