Thank you for Subscribing to Life Science Review Weekly Brief
FREMONT, CA: Machine Learning plays an important role in the healthcare industry. Before heading to the role of ML in the healthcare industry, we need to know the basic difference between classical statistics and machine learning. The major difference between them is their purpose. ML models are designed to make the most accurate predictions possible whereas statistical models are designed for inference about the relationships between variables. Statistics is the mathematical study of data. We cannot do statistics unless we have data. A Statistical model is a model for the data that is used either to infer something about relationships within the data or to create a model that is able to predict future values. Classical statistics always starts with a hypothesis, and generally assumes that some structural relationship exists in the data. When comparing ML to classical statistics, we often look to the assumptions about the data required for the analyses to function reliably. Classical statistical methods typically require the data to have certain characteristics and often use only a few features (called covariates or predictors) to produce results, while ML models might use hundreds or even thousands of parameters in a computer-based method to find similarities and patterns among data.
Machine learning Vs. Statistical modelling
Machine learning is all about predictions, supervised learning, unsupervised learning, etc whereas statistics is about sample, population, hypothesis, etc. Robert Tibshirani, a statistician and machine learning expert at Stanford, calls machine learning “glorified statistics.” Both Statistics and machine learning share the same goal. Machine learning is a branch of computer science and artificial intelligence. It deals with building systems that can learn from data, instead of explicitly programmed instructions. A statistical model, on the other hand, is a subfield of mathematics. Machine learning is comparatively a new field. With the help of cheap computing power and availability of data, scientists could train computers to learn by analyzing data. Whereas, statistical modeling existed long before computers were invented. The major difference between the two is that machine learning emphasizes optimization and performance over inference which is what statistics is concerned about. Machine learning requires no prior assumptions about the underlying relationship between the variables. In contrast, statisticians must understand how the data was collected, statistical properties of the estimator, the underlying distribution of the population they study and kinds of properties that are expected from as many experiments. Machine learning techniques are applied to high dimensional data sets whereas statistical modeling techniques are usually applied to low dimensional data sets.
Machine learning in healthcare industry
ML methods have been shown to consistently deliver outcomes in less time than conventional assessments. For example ML could show a patient’s risk for stroke, coronary artery disease, and kidney failure based on the last 50 blood pressure readings, lab test results, race, gender, family history and latest trial data. It helps clinicians to make better decisions about patient diagnosis and treatment options, while understanding the possible outcomes and cost for each one.
The value of ML is its ability to process huge data sets beyond the scope of human capability and then reliably convert analysis of that data into clinical insights that aid physicians in planning and providing care which ultimately leads to better outcomes, lower costs of care and increased patient satisfaction. ML algorithms can provide immediate benefit to disciplines with processes that are standardized. ML can be trained to look at images,, identify abnormalities, and point to areas that need attention, thus improving the accuracy of all processes.
Classical statistics applications
Classical statistics uses probability theory and underlying distributions, and is usually applied:
1. To low-dimension problems, those with a limited number of potential covariates, predictors, studied populations, or with smaller sample sizes.
2. When you need to know more about data and the properties of predictors to make accurate inferences about the population under study.
3. When you have more structured and complete datasets.
4. When you want to create a scientifically reliable sample dataset from a population in order to conduct valid inferences and draw unbiased conclusions.
In the life sciences industry, the use of classical statistical methods is the foundation for R&D activities and peer-reviewed, real world publications. Statistical analysis plans in this discipline adhere to predefined industry standards. Such cases include randomized clinical trial analysis and patient analytics, such as survival analysis to compare persistence metrics across multiple groups.
Machine learning is more exploratory and less dependent on prior hypotheses or assumptions. Machine learning algorithms are typically far more complex than their statistical counterparts and often require design decisions to be made before an iterative training process begins. It requires no prior assumptions about the underlying relationships between population variables and distributions.
How can ML technology be effectively deployed?
ML technology can only be deployed when organizations have a ML strategy with all the necessary elements: Access to diverse industry data sets and subject-matter expertise, advanced ML technologies allowing for efficient delivery of proofs of concept and solutions and technical expertise to build ML algorithms.
Classical statistics and machine learning need to co-exist; the use of one versus the other should be based on the analytical problem at hand. although both the ML and statistical modeling seem to be different branches, the difference between the two has significantly reduced in the past decade. Both the branches have learned a lot from one another and there will be strong connections between them in the future. It is important to determine which method is effective and appropriate for any given business situation. Effective use of ML in business can demonstrate the value of their products, craft massaging, and execute sales strategies that can deliver commercial success.