Living in the age of Machine Learning : Promising or overhyped..??
In this ‘ Online Era ’, the identity and even personality of each free spirit can be reckoned from their engagement in social platforms. “How come technology hold much from us ?” This question leads us to admire the remarkable mystery of prevailing technologies; peculiarly Artificial Intelligence (AI), Machine Learning (ML) and Deep Learning (DL). Even though the journey of Artificial Intelligence has started around 1950’s, the full-fledged boom was gained with the emergence of Machine learning, in 1980’s, which got further advanced through the contributions of Deep Learning. The promising technology of Machine Learning has brought an astounding breakthrough in the field of industry and business.
With the advent of Machine Learning, the most awaited ‘Tomorrowland’ of computers and electronic gadgets-completing tasks with no pre-designed algorithms- was achieved. To inculcate such an ability to a machine, data from various sources — with required attributes — should be available to train the machine. Apart from these pieces of training, machines can now accumulate older experience in order to make decisions, with the aid of special algorithms. As per studies conducted by the Bank of America Merill Lynch, “Over the next five years, the market will tend to $153 billion compared to $58 billion in 2014”. Followed by that the market share of ML applications, over $2 billion when compared with other fields in Artificial Intelligence.
According to the life cycle assessment curve, the spotlight over the field of business will fade soon, as it enters the phase of saturation. And the next role will be played by the healthcare industry. The researches from health care data science have gained a greater attention than ever before, after the prediction of a likely growth in healthcare by the economists. These predictions have indeed accelerated the study on various practical applications of ML in healthcare.
Earlier, among the constraints that hinder the reshaping of the healthcare sector, the most contributing was the unavailability of big data. But this scenario has changed with the emergence of biomedical equipment with very high computing power and storage facilities, making the medical field a sprouting bed for ML.
Considering the advancements in the medical field with ML, the most popular are the cancer prognosis, development of learning aid for differently abled students, optimising clinical tests, diagnoses of diseases, geriatric aids, predicting personalized health outcomes. Among these, a wide scope of study and opportunities lie within the field of predicting personalized health outcomes.
A wide scope of study and opportunities lie within the field of predicting personalised health outcomes.
Perplexities regarding choosing the right deserving area soon got resolved with the insight from Dr.Neetha George, Gynaecologist Jubilee Mission Hospital Thrissur, who informed about the alarmingly leading rate of PCOS (Polycystic Ovary Syndrome) patients in the existing female population. PCOS — an endocrine disorder with abnormalities in the rate of hormones and elevated levels of male-hormone- seen among one-in-ten women of existing female population within 18–44 years of reproductive age.
The major signs of PCOS are obesity, acne, absence of menstrual cycle, excessive hair growth etc. Women with PCOS have higher rates of endometrial cancer (cancer of inner layer of the uterus), cardiovascular disease, dyslipidemia(irregularity in blood lipid level), type-2 diabetes mellitus and infertility. Polycystic ovary syndrome represents 80% of anovulatory (condition with no ovulation) infertility cases. Among all the conditions infertility is the most dreaded menace over mankind. The Chairman of KARE ( Krishna Assisted Reproduction and Endoscopy Centre) Centre, Dr.Krishnankutty shared his opinion during the consultation, “ Emerging technologies that help the doctors to empathize with the psychological trauma suffered by the childless couples are always welcoming ”. This paved way for the finalization of the project topic, ‘An early detection and prediction system for PCOS patients using machine learning algorithms’. Medical diagnosis of PCOS during its initial stage would help to gain control over it to a greater extent.
‘An early detection and prediction system for PCOS patients using machine learning algorithms’
Big data ML allows aggregating the information from different sources (physical and biochemical parameters from hospitals and clinics) and give predictions in real time after several data processing and feature extraction.The promising features that contribute the most towards PCOS condition will be identified and classified with more weightage than other parameters. By analysing the cumulative weights, automatic screening of PCOS patients with non PCOS patients could be done. Based on the PCOS status, customised outputs are given to patients through a patient-mobile interface. As a future scope we would like to use the big data available for the infertility prediction of PCOS patients.
Major challenges faced while using machine learning includes -risk in handling confidentiality of the patient data, demand of enormous data, machines with high computing power, requirement of large processing time, applications are not well defined.
Machine learning is a promising evolution and big companies are bet on this technology.
References:
Ndefo, U. A., Eaton, A., & Green, M. R. (2013). Polycystic ovary syndrome: a review of treatment options with a focus on pharmacological approaches. P & T : a peer-reviewed journal for formulary management, 38(6), 336–55.
Melo, A. S., Ferriani, R. A., & Navarro, P. A. (2015). Treatment of infertility in women with polycystic ovary syndrome: approach to clinical practice. Clinics (Sao Paulo, Brazil), 70(11), 765–9.
http://www.quickmba.com/marketing/product/lifecycle/
https://theappsolutions.com/blog/development/machine-learning-in-mobile-app/