Skip to main content


Analyzing Net Promoter Score using Machine Learning

Overview This article presents a reference implementation of a Net Promoter score analysis project that is built by using Machine Learning. In this article, we will discuss associated generic models for holistically solving the problem of industrial Net Promoter score. We also measure the accuracy of models that are built by using Machine Learning and assess directions for further development. Net Promoter Score The Net Promoter Score is an index ranging from 0 to 10 that measures the customer’s loyalty to the brand. Calculation Customers are surveyed on one single question.  “On a scale of 0 to 10, how likely are you to recommend this company’s product or service to a friend or a colleague?” Based on their rating, customers are then classified into 3 categories: detractors, passives, and promoters.

Recent posts

SQL Table-Valued Function to Find the Best Trend Line Using Linear Regression

Linear regression is probably one of the first regression techniques we run across. It is useful for explaining (and perhaps forecasting) data that has a linear nature, but where the underlying process is unknown. 

This solution focuses on the practical aspects of finding the best trend line for a financial time series using SQL function.
If you have a large dataset with multiple sets of x and y data points associated with some attribute, you can do a linear regression on each set with one query using GroupID.

The great strength of the linear regression line is that it can clarify sometimes ambiguous line charts and quantify the results via the slope of the line.

Building a ML classifier on dentistry payment plans data using C#

I prefer the Naive Bayes approach, because while having small set of data, it seems to make better decisions than I did in many cases. Also, the other type of classifier seems to do better on clear-cut cases, but doesn't seem to handle uncertainty so well. In addition,  the naive Bayes classifier corresponds to a linear classifier in a particular feature space.

Main Program
Text Pre-processing Class Naive Bayes Class


This blog presents a novel non-invasive approach for early diagnosis of prostate cancer from dynamic contrast enhanced magnetic resonance imaging (DCE-MRI). 
In order to precisely analyze the complex 3D DCE-MRI of the prostate, a novel processing frame work that consists of four main steps is proposed. The first step is to isolate the prostate from the surrounding anatomical structures based on a Maximum a Posteriori (MAP) estimate of a log likelihood function that accounts for the shape priori, the spatial interaction, and the current appearance of the prostate tissues and its background (surrounding anatomical structures).
In the second step, a non-rigid registration algorithm is employed to account for any local deformation that could occur in the prostate during the scanning process due to the patient’s breathing and local motion. 
In the third step, the perfusion curves that show propagation of the contrast agent into the tissue are obtained from the segmented prostate of the wh…

What is Mean to Know The Patient Life Time Of Your Dental Patients?

How to Measure Patient Lifetime Value
The Patient Lifetime Value is the key success factor in operating any dental practice and it shows the ability to retain patients once they come in the door. By focusing on this long-term value, the dental practice can enjoy higher returns from his existing client base.

What Is Dental Patient Life Time Value?PLT value is a prediction of the total worth of a dental patient to a dentistry over the entirety of their relationship.
Most often, a dentist starts looking for advertising when he or she notices an acute need for patients, hoping for a short-term influx of new work. If the advertising is successful in bringing in patients, it is often terminated – at least until the next time the practice is not attracting enough new patients. Similarly, new patients are deemed monetarily unrewarding if their service includes only standard cleaning and examination procedures. This lack of long-term focus is perpetuated by the absence of any meaningful material …

3D Medical Image Segmentation

This work is based on a Maximum a Posteriori (MAP) estimate of a log-likelihood function that accounts for the shape priori, the spatial interaction, and the current appearance of the tissues and its background (surrounding anatomical structures).

 Main references:
Novel stochastic framework for accurate segmentation of prostate in dynamic contrast enhanced MRI International Workshop on Prostate Cancer Imaging, 121-130A new 3D automatic segmentation framework for accurate segmentation of prostate from DCE-MRI 2011 IEEE International Symposium on Biomedical Imaging: From Nano to Macro

Store server application for dicom images

Store Server Program Using C++ and DICOM Network Protocol
A storage server program is able to receive dicom images from a source AE such as console, store them on secure user account, which matches the key attribute of the image. Create scan sheet of the images that have been received, parse out important header information, and store that header information in a database.
SYNOPSIS  It starts automatically at boot-in and running on port 104 in the background. If you don’t want the server running on port 104, edit store_server and run it on a new terminal window. ./store_server (no arguments)
To change port number you have to edit the port number as shown below.   store_server (Port number, './Temp', './handlerdb', 'store_server_config.cfg', 1); DESCRIPTIONAfter an images are received from a source AE, the store program will run another script (handlerdb). This script will be responsible for moving the images into the user account with a specific organization of t…