Friday, November 30, 2018

Insurance Claims Data Analysis and Analytics

This Insurance Claims Data Analysis Dashboard includes Motor Insurance Claims Data based in the UK. The application runs on Qlik Sense Associative Engine which allows users to perform in depth analysis of the claims payments across a wide range of factors including time, location and claim type. In the application you can see how poor data quality and significant outliers can have a direct impact on the performance results of the company.

The Insurance Claims Analytics video below shows how you can use business intelligence to analyze insurance claims data to identify claims fraud, unusual transactions and data quality issues. You can try the Insurance Claims Data Analysis Dashboard yourself here in the demo page.

One of the issues insurance companies face is fraud. Fraud attempts have seen a drastic increase in recent years with the increase in online businesses thus making fraud detection more important than ever. Despite efforts on the part of the affected institutions, hundreds of millions of dollars are lost to fraud each year and quite likely to increase as well. Just like a needle in a haystack relatively few cases show fraud in a large population. Finding these is not just tricky but sometimes impossible too.

A key weapon for insurers in identifying these fraud perpetrators is the analysis of data. In a classical data analysis scenario, insurers need to be able to search for associations in data between similar types of claims, in similar locations, including something unique like a mobile phone number. These associations between the data can lead to a significant increase in identifying the groups of people that commit these types of fraud. This is exactly where a data visualization solution like Qlik Sense can play an important role in this activity.  Qlik Sense can help Insurance Fraud Analysts identify trends, patterns and examples of fraudulent Whiplash claims.

One step further can be predicting which claims are fraud cases using predictive analytics. Predictive analytics do not require insurers to go through the relationships in their data manually and try to find out the cases where fraud probability is high. This task can be tedious if there are many parameters in the claims data but can easily be handled by a predictive model.

For example below, you can see how an automated machine learning tool (Enhencer in the below case) can help to identify fraud cases.


Monday, October 15, 2018

What is Automated Machine Learning (AutoML)?

Machine learning is a field of artificial intelligence that uses statistical techniques to give computer systems the ability to "learn" (e.g., progressively improve performance on a specific task) from data, without being explicitly programmed. This trained machine learning model can later be used to predict the probability of a future event within an acceptable reliability.

For example, suppose that you want to predict the customer churn for a customer. You can use your historic customer data with customers you kept and lost to train a model using machine learning techniques. Later, you can feed a new customer info (or a set of new customer data) to the model and predict their probability to be lost.

This is a two step process. First you train a model and then you use this trained model to predict (or score in machine learning terms) new data.

Machine Learning training and scoring
Although it looks straightforward, traditional machine learning process requires some intermediary steps.  A typical machine learning application, practitioners must apply the appropriate data pre-processing, feature engineering, feature extraction, and feature selection methods that make the dataset amenable for machine learning. Following those preprocessing steps, practitioners must then perform algorithm selection and hyperparameter optimization to maximize the predictive performance of their final machine learning model.

Traditional machine learning process.
Many of these steps are often beyond the abilities of non-experts and automated machine learning (AutoML) addresses this problem. It is the process of automating the end-to-end process of applying machine learning to real-world problems.

As Janakiram MSV writes in Forbes (Why AutoML Is Set To Become The Future Of Artificial Intelligence):
“AutoML focuses on two aspects – Data acquisition and prediction. All the steps that take place in between these two phases will be abstracted by the AutoML platform. Essentially, users bring their own dataset, identify the labels, and push a button to generate a thoroughly trained and optimized model that's ready to predict.”

Automated Machine Learning
For example, take a look at the predictive fraud analytics for insurance sector example below. The AutoML application used here is cloud based Enhencer Predictive Story Teller. After integrating claims data to the predictive system, predictive model creation is done with a few clicks without expertise in data science and statistics (you can watch a more detailed introduction and demo here : Predictive Analytics and Machine Learning Introduction - Customer Propensity Example)


Thursday, October 11, 2018

Predictive Analytics and Machine Learning Webinar - Customer Propensity Example

Beyond business intelligence, where you can slice and dice your historic data to understand what happened and why, predictive analytics will tell you what will happen and make powerful predictions about the future. This will replace intuition based decisions with data driven decision making process. You will also see how powerful machine learning algorithms are open to business users now thanks to self service advanced data analytics tools like Enhencer.

 Enhencer enables business users to use advanced analytics and machine learning on their data without data science knowledge. This allows them to create predictive models on certain metrics within minutes and then predict individual records based on the model (model can be exported to SQL or your favorite BI solution like Microsoft Power BI, Tableau or Qlik).


 For more information please visit https://www.enhencer.com/

Monday, September 10, 2018

Predictive Fraud Analysis - Enhencer (self service data analytics)

Fraud attempts have seen a drastic increase in recent years with the increase in online businesses thus making fraud detection more important than ever. Despite efforts on the part of the affected institutions, hundreds of millions of dollars are lost to fraud each year and quite likely to increase as well. Just like a needle in a haystack relatively few cases show fraud in a large population. Finding these is not just tricky but sometimes impossible too.

Predictive analytics can make fraud detection very easy and help you to seek out the needle in the haystack in no time.
  • Using the dynamic segmentation, you can find the customer segments with the highest fraud rates.
  • Predict the fraud activity
  • You can learn the features of the customer segment likely to become a fraud and take actions to reduce the frauds.



Fast & Powerful Data Analysis

You don't need expensive and complex data science resources to benefit from advanced data analytics. Enhencer enables you to import, analyze and predict in less than few minutes requiring no coding or statistical expertise at all. Enhencer puts the data analysis in a nutshell with 4 steps.

1 - Connect/Upload Data
Upload data of many formats or Import data from various sources like Typeform, Survey Monkey, Google Form and many more in a matter of seconds.

2 - Instant Actionable Insights
Bring down the analysis time from days to minutes as Enhencer empowers the data analysis with decision tree algorithm in an accessible manner for everyone.

3 - Powerful Segmentation
Enhencer is designed with machine learning algorithms that can explore and provide the best & reliable segments automatically no matter the complexity of the data.

4 - Predictive Models
Enhencer builds the perfect predictive model & digs out the variable that truly effects your target automatically leaving you the task of just clicking the predict button.

Predictive analytics segmentation
Advanced data analytics segmentation
Dynamic Data Analytics Features

Enhencer brings all these features together to present the most complete data analysis package featuring from simple & stunning visuals to the most advanced data mining & predictive modeling algorithms. Enhencer also brings machine learning algorithms to the table making the data insights and predictions more accurate and reliable than ever.

Machine Learning Algorithms
Enhencer takes an approach for improving performance and reliability by using machine learning algorithms.

Powerful Segmentation
Enhencer provides powerful & reliable segments automatically from the data using machine learning algorithms.

Stunning Visual Stories
Using intuitive and interactive interface of Enhencer acquire stunning visual stories for all your data types.

Robust Prediction
Predict the unknown using powerful predictive models such as decision tree, random forest, XGBoost & neural network.

Likert & Promoter Score
Survey friendly features to spot the business growth opportunities that no other data analysis packages provide.

Flexible Integrations
Connect and integrate the different data sources from online data collection platforms and database platforms.


For more information, demo-request or free trial, please visit www.enhencer.com

Predictive Insurance Data Analytics and Segmentation - Enhencer



Fast & Powerful Data Analysis

You don't need expensive and complex data science resources to benefit from advanced data analytics. Enhencer enables you to import, analyze and predict in less than few minutes requiring no coding or statistical expertise at all. Enhencer puts the data analysis in a nutshell with 4 steps.

1 - Connect/Upload Data
Upload data of many formats or Import data from various sources like Typeform, Survey Monkey, Google Form and many more in a matter of seconds.

2 - Instant Actionable Insights
Bring down the analysis time from days to minutes as Enhencer empowers the data analysis with decision tree algorithm in an accessible manner for everyone.

3 - Powerful Segmentation
Enhencer is designed with machine learning algorithms that can explore and provide the best & reliable segments automatically no matter the complexity of the data.

4 - Predictive Models
Enhencer builds the perfect predictive model & digs out the variable that truly effects your target automatically leaving you the task of just clicking the predict button.

Predictive analytics segmentation
Advanced data analytics segmentation
Dynamic Data Analytics Features

Enhencer brings all these features together to present the most complete data analysis package featuring from simple & stunning visuals to the most advanced data mining & predictive modeling algorithms. Enhencer also brings machine learning algorithms to the table making the data insights and predictions more accurate and reliable than ever.

Machine Learning Algorithms
Enhencer takes an approach for improving performance and reliability by using machine learning algorithms.

Powerful Segmentation
Enhencer provides powerful & reliable segments automatically from the data using machine learning algorithms.

Stunning Visual Stories
Using intuitive and interactive interface of Enhencer acquire stunning visual stories for all your data types.

Robust Prediction
Predict the unknown using powerful predictive models such as decision tree, random forest, XGBoost & neural network.

Likert & Promoter Score
Survey friendly features to spot the business growth opportunities that no other data analysis packages provide.

Flexible Integrations
Connect and integrate the different data sources from online data collection platforms and database platforms.


For more information, demo-request or free trial, please click.




Friday, September 7, 2018

Self Service Predictive Analytics and Segmentation - Enhencer



Fast & Powerful Data Analysis

You don't need expensive and complex data science resources to benefit from advanced data analytics. Enhencer enables you to import, analyze and predict in less than few minutes requiring no coding or statistical expertise at all. Enhencer puts the data analysis in a nutshell with 4 steps.

1 - Connect/Upload Data
Upload data of many formats or Import data from various sources like Typeform, Survey Monkey, Google Form and many more in a matter of seconds.

2 - Instant Actionable Insights
Bring down the analysis time from days to minutes as Enhencer empowers the data analysis with decision tree algorithm in an accessible manner for everyone.

3 - Powerful Segmentation
Enhencer is designed with machine learning algorithms that can explore and provide the best & reliable segments automatically no matter the complexity of the data.

4 - Predictive Models
Enhencer builds the perfect predictive model & digs out the variable that truly effects your target automatically leaving you the task of just clicking the predict button.

Predictive analytics segmentation
Advanced data analytics segmentation
Dynamic Data Analytics Features

Enhencer brings all these features together to present the most complete data analysis package featuring from simple & stunning visuals to the most advanced data mining & predictive modeling algorithms. Enhencer also brings machine learning algorithms to the table making the data insights and predictions more accurate and reliable than ever.

Machine Learning Algorithms
Enhencer takes an approach for improving performance and reliability by using machine learning algorithms.

Powerful Segmentation
Enhencer provides powerful & reliable segments automatically from the data using machine learning algorithms.

Stunning Visual Stories
Using intuitive and interactive interface of Enhencer acquire stunning visual stories for all your data types.

Robust Prediction
Predict the unknown using powerful predictive models such as decision tree, random forest, XGBoost & neural network.

Likert & Promoter Score
Survey friendly features to spot the business growth opportunities that no other data analysis packages provide.

Flexible Integrations
Connect and integrate the different data sources from online data collection platforms and database platforms.

sssss


For more information, demo-request or free trial, please click.





Wednesday, May 2, 2018

Enhencer Predictive Storyteller - Data Analytics Live Demo

Enhencer is a visual data analytics and predictive intelligence platform focuses on the business intelligence and data visualization concepts after the emergence and popularity of the machine learning methods. With Enhencer You no longer need out-of-fashion summary tables and plots, prediction algorithms those are not always accurate, and bulky software to manage prediction processes. Enhencer generates predictive functions that you can use in your action plans by clearly showing the causalities in the data. Enhencer is also an analytical presentation tool which provides you with an elegant and interactive dashboard to display reports.

There is a live demo of Enhencer online you can try on your own. If you have any question or a request for demo, you can fill the request for a demo form here or email Enhencer. For more info, you can also refer to Enhencer Predictive Intelligence Online Help.

# Open Enhencer's Customer Satisfaction Data Analytics live demo.

Enhencer screen is divided into 3 panels.

Left Panel in the analyze screen contains all the variables from the data. The green highlighted variable indicated the particular variable is selected as “Target Variable” for the analyzes. You can simply change the target variable by simply clicking a different variable.

Currently, Customer Status variable is selected. This variable is categorical and has two possible values : Retention and Churn. As you can see, overall customer retention is 54.2% and churn is 45.9%.



In order to obtain deeper insights from the data, users need to use the “Enhance” task. Behind the curtain Enhencer used Machine Learning Algorithms to fetch the deeper insights from the data.

Enhance task can be performed by clicking the “Enhance” button. Enhance lets you to use the machine learning algorithm to dig out the dapper insights from the data.

# Click Enhance button.

Upon clicking the “Enhance” button a new window will pop up named “Select”. You need to select the relevant variables to observe the effects of the selected variables on the target variable.

Blue highlighted variables are the ones that are currently selected for the Enhance task. The deselected variables are represented with the non-highlighted ones.

#Click ALL in Select window to select all variables. Click OK at the bottom of Select Window.




Right side of the analyze screen is reserved for hierarchy view. When “Enhance” is active the tree view is presented to the users. Here Relations and effects of other variables with the target variable can be observed.

Leaf represents a segment and the features above the leaf represents the path to that segment.



Tree View branches from the most important variable effecting customer status to least. Currently, overall is selected which is 4,000 data point size.  The first leaf shows the online usage and the split is 14.9895. As you can see, 26.1% of the customers have less than 14.9 hours online usage and their churn is the highest at 84.9%.



# Select Retention from Target Choice to let Enhencer perform segmentation.


Enhencer will create the segments with the highest lift at the top.

Segments is a powerful tool that shows segments of the target variable and their features. When “Enhance” is active segments can be accessed by simply clicking the “Segments” button. If target variable is categorical like this example then in order to access the segments the user needs to choose one of the target choices form the “Target Choice” drop down menu. For any other variable type, it can be accessed directly.

Right side of the analyze screen in this case shows all the segments available for the target variable. Segments are sorted from highest gain to the lowest gain.

# Click Segment 1 to see its properties. As you can see below, this segment has the highest lift.

The proportion of the target choice in the selected segment compared to the overall population is represented by “Gain & Lift” measures. For this instance, in segment number 1 among 271 customers 90.8% of the customers are retained which is 1.68 times (67.6%) higher than the overall population with 54.2% retention rate.

In Segment 1, you can see the highest retention segment. These are people with 14.9 hours or more online usage, primary social network Facebook, and so on.



You can see which variables effect the target variable most using Effects View.

# Click Effects button.

Effects is a powerful tool that shows the variables that effects the target variable the most. This means among hundreds of different factors from the data these are the ones to affect the target variable the most.

When “Enhance” is active effects can be accessed by simply clicking the “Effects” button. The right side of the screen shows the effects of other variables. The variables are sorted from “High Dependency” to the “Low Dependency”.


When “Effects” is active users can also take benefit of prediction task. In Enhencer you can either predict by simulation or by uploading a new dataset.

In order to take benefit of the “Predict” task user must have [A] “Enhance” active and be on the [B] “Effects” tab.

Effects is a powerful tool that shows the variables that effects the target variable the most. This means among many different factors from the data these are the ones to affect the target variable the most. Moreover, these are the variables that the user can use to predict new cases either by simulating a case from the variable list in the right or uploading a completely new data.

If the target variable is categorical as in this case, Target Choice should be selected. In our example, Retention is already selected.

Prediction can be of two types. One is the classical test data method, where prediction scores are observed for a new dataset after the model is trained. Another is to simulate a very specific case just to observe how the data scores might change or how specific variables contributes to the prediction scores. This example is the latter one, simulating a certain case.

# Select Online Usage above 20 hours and Facebook as the only primary social media account. Click Predict.


Enhencer predicts that you will have 75% chance to retain this type of users.