Storage Vendor News
Hu Yoshidas blog
Machine Learning Helps to Control Diabetes
Mon, 19 Mar 2018
There is a good probability that you or someone you know has diabetes. The World Health organization believes that an estimated 8.8 percent of the adult population worldwide had diabetes. This figure is projected to rise to 9.9 percent by the year 2045. Type-2 diabetes is the most prevalent form of diabetes and affects more people as the population ages. Today one in every four Americans, 65 years or older has Type-2 diabetes. The spread of Western lifestyles and diet to developing countries has also resulted in a substantial increase.
Diabetes is a chronic, incurable disease that occurs when the body doesn't produce any or enough insulin, leading to an excess of sugar in the blood. Diabetes can cause serious health complications including heart disease, blindness, kidney failure, and lower-extremity amputations and is the seventh leading cause of death in the United States. Diabetes can be controlled by medication, a healthy diet, and exercise.
The problem with medication is that there are many ways to treat the disease with different combinations of drugs. Some medications breakdown starches and sugars, other decreases the sugar your liver makes, some affect rhythms in your body and prevent insulin resistance, others help the body make insulin, still others control how much insulin your body uses, some prevent the kidneys from holding on to glucose, and others help fat cells to use insulin better. New medications are being developed continuously as the population of diabetics increases. Diabetics often need to take other medications to treat conditions that are common with diabetes like heart health, high cholesterol, retinopathy, and high blood pressure. The efficacy of the drugs changes with age and other physical factors. There are also different side effects depending of the individual’s situation, and the drugs can be expensive. The effectiveness of the treatment is measured every three months by a blood test for a measure called A1C. A1C measures the average blood glucose level for the last 3 months. An A1C measure of 7.0% indicates that the blood glucose level and the diabetes are under control. However, 7.0% is an ideal reading and higher readings may be acceptable depending on the individual. Up to now the prescription of medication is usually a trial and error approach and more than half of diabetes patients fail to achieve the treatment targets according to the World Journal of Diabetes. The selection and monitoring of the most effective medication or combination of medications that is also safe, economical and better tolerated by patients is often hit or miss.
On March 12, Hitachi and the University of Utah Health, a leading institution in electronic health records and interoperability clinical information systems research announced the joint development of a decision support system that allows clinicians and patients to choose pharmaceutical options for treating type-2 diabetes. The system uses machine learning methods to predict the probability of a given medication regimen achieving targeted results by integrating with electronic health records which allows for guidance that is personalized for individual characteristics.
The system compares medication regimens side-by-side, predicting efficacy, risks of side effects, and costs in a way that it is easy for clinicians and patients to understand.
Using Machine learning combined with the individual’s individual health records will increase the probability of selecting the right combination of medications that will help individuals reach their targeted goals to control diabetes. Think how this could be applied to other treatments like chemo therapy for cancer. If you know anyone with diabetes please forward this post to them so they can understand what is possible when you apply machine learning to the control of this disease.
Last week Hitachi Vantara Labs announced Machine Learning Model Management To accelerate model deployment and reduce business risk. This innovation provides machine learning orchestration to help data scientists monitor, test, retrain and redeploy supervised models in production. These new tools can be used in a data pipeline built in Pentaho to help improve business outcomes and reduce risk by making it easier to update models in response to continual change. Improved transparency gives people inside organizations better insights and confidence in their algorithms. Hitachi Vantara Labs is making machine learning model management available as a plug-in through the Pentaho Marketplace.
Machine learning explores the study and construction of algorithms that can “learn” from and make predictions on data through building a model from sample inputs without being specifically programmed. These algorithms and models become a key competitive advantage – and potentially a risk. Once a model is in production, it must be monitored, tested and retrained continually in response to changing conditions, then redeployed. Today this work involves considerable manual effort and is often done infrequently. When this happens, prediction accuracy will deteriorate and impact the profitability of data-driven businesses.
David Menninger, SVP & Research Director, Hitachi Vantara Research, said, “According to our research, two-thirds of organizations do not have an automated process to seamlessly update their predictive analytics models. As a result, less than one-quarter of machine learning models are updated daily, approximately one-third are updated weekly and just over half are updated monthly. Out-of-date models can create significant risk to organizations.”
So, what is Machine Learning Model Management and where does it fit in the analytic process?
Machine Learning Model Management recognizes that machine learning models need to be updated periodically as the underlying distribution of data changes and the model predictions become less accurate over time. The four steps to Machine Learning Model Management include, Monitor, Evaluate, Compare, and Rebuild as shown in the diagram above. Each step implements a concept called “Champion/Challenger”. The idea is to compare two or more models against each other and promote the one model that performs the best. Each model may be trained differently, or use different algorithms, but all run against the same data. These 4 steps to Machine Learning Model Management is a continuous process and can be run on a scheduled basis to reduce the manual effort of rebuilding these models.
Hitachi Vantara’s implementation of Machine Learning Model Management is part of the Pentaho data flow which makes machine learning easier by combining it with Pentaho’s data integration tool. In the diagram above the preparation of data may take 80% of the time to implement a model with preparation processes that rely on coding or scripting by a developer. Pentaho Data Integration empowers data analysts to prepare the data they need in a self-service fashion without waiting on IT. An easy to use graphical interface simplifies the transformation, blending, and cleansing of any data for data analysts, business analysts, data scientists, and other users. PDI also has a new capability that provides direct access to various supervised machine learning algorithms as full PDI steps that can be designed directly into your PDI data flow transformations.
For more information on PDI and how it integrates with Machine Learning Model Management see the following blog posts by Ken Wood.