Hi, I am a Statistician with PhD in Statistics from Calcutta University. I have been working in IT Industry for past 20 years. I love thinking independently. Regards.
Machine generated data as recorded by the machine sensors is studied to detect possible outliers.... more Machine generated data as recorded by the machine sensors is studied to detect possible outliers. Mahalanobis Distance as it follows F Distribution under normality assumptions, is used to detect the outliers. A near-real-time dashboard plotting the outlier percentages over a moving window in time is proposed to monitor the health of the machine. Optimal stop for maintenance is suggested based on 'Secretary Problem'.
The revenue of a business was to be predicted ahead of time, say by months. Keeping the revenue r... more The revenue of a business was to be predicted ahead of time, say by months. Keeping the revenue records as the target variable, several related explanatory variables were selected, data of which were collected from internet public domain. Best lag time ranges were calculated and multiple linear regression were performed for t. Minimum of the best lags was achieved as the lead time for the revenue prediction.
Considered is a finite regular grid of squares. The main question addressed is on the minimum num... more Considered is a finite regular grid of squares. The main question addressed is on the minimum number of edges needed to be removed to make the grid square-free.
Machine generated data as recorded by the machine sensors is studied to detect possible outliers.... more Machine generated data as recorded by the machine sensors is studied to detect possible outliers. Mahalanobis Distance as it follows F Distribution under normality assumptions, is used to detect the outliers. A near-real-time dashboard plotting the outlier percentages over a moving window in time is proposed to monitor the health of the machine. Optimal stop for maintenance is suggested based on 'Secretary Problem'.
The revenue of a business was to be predicted ahead of time, say by months. Keeping the revenue r... more The revenue of a business was to be predicted ahead of time, say by months. Keeping the revenue records as the target variable, several related explanatory variables were selected, data of which were collected from internet public domain. Best lag time ranges were calculated and multiple linear regression were performed for t. Minimum of the best lags was achieved as the lead time for the revenue prediction.
Considered is a finite regular grid of squares. The main question addressed is on the minimum num... more Considered is a finite regular grid of squares. The main question addressed is on the minimum number of edges needed to be removed to make the grid square-free.
Uploads
Papers by Moloy De
Drafts by Moloy De