top of page


Author: - Tanya Juneja

Machine learning plays a vital role in healthcare. It yields better results in health care domain. As per McKinsey report machine learning and big data in pharmacy and medicine could generate revenue up to $100B annually .

This is due to fact the following facts:-

  • The faster decision-making,

  • Improved efficiency during clinical trials,

  • Optimized innovation.

Classification of ML in healthcare is as follow:-

  1. Disease Identification/Diagnosis

  2. Personalized Treatment/Behavioral Modification

  3. Drug Discovery/Manufacturing

  4. Clinical Trial Research

  5. Radiology and Radiotherapy

  6. Smart Electronic Health Records

  7. Epidemic Outbreak Prediction

Machine learning can be applied to health care data to develop robust risk models. As the ratio of doctor to patients in India is 1:1700 which is far higher than the recommended ratio of 1 in every 1000 patients by WHO which is quite overburdened. The spontaneous increase of efficient healthcare providers is not possible as it is time consuming. Use of machine learning and artificial intelligence technologies can enhance the productivity and precision of existing ones. Use of these technologies will help in serving more patients in a less time and also improve healthcare outcomes and also reduce the healthcare expense.

Both supervised and unsupervised learning can be used to design a model and we can choose the best fit one.

Fig1:-Machine learning classification

Predictive analytics

It is a type of analysis in which pattern are found out from variety of data mining, predictive modeling and machine learning. Predictive models works on patterns found in historical and transactional data for identification of risks and opportunities. Models allows to capture relationships among many attributes which will allow assessment of risk or potential associated with a particular set of conditions, guiding decision-making for the upcoming event or future possibility.

Predictive analysis in healthcare

The significance of predictive modeling in healthcare can easily be found in emergency care, surgery and intensive care, where the outcome of a patient is directly related to the quick reaction and acute decision making of the care provider or doctor when there is an occurrence of unexpected situation turn for the worse conditions. But not all predictive analytics in healthcare require an experienced team to make thing work.

Predictive modeling in healthcare can help an organizations identify patients who have increased risks of developing chronic conditions early in the disease’s development stage , also helping patients to avoid costly and difficult to treat methods for the worsen health problems. Creating a database based on health conditions, as well as demographic factors such as Medicaid and disability status, gender, age, and whether a beneficiary lives in the community or in an institution can give healthcare data companies insight into which individuals could benefit from personalized healthcare or wellness programs services.

The management of high-risk patients is essential for improving quality in an organization. With the use of predictive analytics healthcare companies can proactively identify patients who are at highest risk of poor health and could benefit the most from this process.

Patients face possible threats to their well being while still in the hospital, including the development of difficult-to-treat infections, or abrupt downturns due to their existing conditions. Predictive modeling in healthcare can help doctors tale action as quickly as possible to changes in a patient’s vitals, and make it easier to identify an upcoming worse of symptoms before that already takes place while making a perfect action plan to save life.

COVID-19 in India: A review

COVID-19 pandemic outbreak has been recent area which is attempted by a wide range of researchers from the very beginning of cases in India. Initial analysis of available models revealed large variations in scope, assumptions, predictions, course, effect of interventions, effect on health-care services, and further.

In an review of the pandemic outbreak The Pearson’s correlation coefficient (r) between predicted and actual values (n = 20) was 0.7 (p = 0.002) with R2 = 0.49. For Susceptible, Infected, Recovered (SIR) and its variant models (n = 16) ‘r’ was 0.65 (p = 0.02).

Mathematical techniques used for modeling were also varied from researchers. These are the technique used by researchers. The below fig 2 shows that most studies (17, 56%) were published using SIR model or its variant. The assumptions made by different models regarding R0 (R Naught), infectious period, recovery time, serial interval.

Fig2:-Type of mathematical modeling used for modeling COVID-19, ∗SIR, susceptible, infectious, recovered; SEIR, susceptible, exposed, infectious, recovered.


Linear regression using python gives the prediction value of 49967.36422545 for India with date and total cases considered as prediction. Similarly, one can take any values for prediction purpose.

· Confusion matrix

Fig 3:- confusion matrix

Graphs and data information:-

Different types of graph help us study the ongoing trend followed by data and help us predict more accurately.

Fig 4:- Bar plot

Fig 4:- Null values

Fig 5:- total cases per day

GitHub Link:-







44 views0 comments

Recent Posts

See All
bottom of page