@Author Anustup Mukherjee
Linkedin: www.linkedin.com/in/anustupmukherjee
COVID19-PREDICTION MODEL ,DATA VISUALIZATION AND PREDICTION CURVE
ACCURACY 95 TO 97 %
The details about the scripts are just discussed below.The model is just a integrated model but for convenience its being divided into three ipnyb scripts. To show the output clearly.Its basically predicting : • Number of Covid-cases in India as a country in upcoming days,months and data visualization • State wise prediction and data visualization • International Prediction and data visualization • Dynamic Indian Prediction Curve to show when this situation will end and upcoming prediction on year ,month and day basis upcoming.
ALL MODELS GET AUTO-UPDATED DAILY AS IT FETCHES DIRECTLY THE DATA FROM ONLINE SOURCE WHEN EVER PERSON RUN THE SOFTWARE.NOTHING TO DO ONLY PRESS RUN ,SET THE DATE IS YOU WANT SPECIFIC PREDICTION OTHER WISE IT WILL GIT PREDICTION,CURVE,DATA VISUALIZATION BY ITS OWN OF COVID 19.ALL ARE ON AI BASED APPROACHES.
The scripts are discussed as follows with accuracy and comparison :
Script1: Statewise_prediction_COVID19:
This is based on stat computing ML Bayesian model with FbProphet model fir best predictions.It shows state wise Covid predictions of India in upcoming month.It get updated automatically as it fetch online data and gets automated trained by adaptive learning . Here we took a state westbengal to predict by the model .All predictions got 97% Accuracy.
Script 2: India COVID19 PRED CURVE AND INTERNATIONAL PRED This script is based on the same above mentioned tech stack .But it predicts with further advancement .Vedio is provided of the visual dynamic pred curve of India with nature,cases and all possible things.Along with this it provides number of upcoming cases in coming days in India as well all over the world : Accuracy 95 to 97%
Script 3: INTERNATIONAL DATA MODELLING AND PRED Its based on a Japanese Kalman filter based on r2py R in Python script model with docker container of jupyter+R+python+h20.io+spark.It is showing international case prediction by visual data approach .Its is also updated daily .