In the big data age, little data is being brought up.

&http://www.aliyun.com/zixun/aggregation/37954.html ">nbsp;

In this month's new issue of CACM, two more articles talk about small data. Large data is very popular, small data can not be ignored, especially for personalized medical care, small data is indispensable. Moreover, the research of small data and large data is closely related.

Large data, smart devices, such as genomics, proteomics, metabolic groups, and so on, are going to change medicine. Another force, however, is to change our thoughts and practices about health, that is, the small data driven by personal digital tracking. Consider a cloud based application that keeps a picture of your health in a continuous, safe, private analysis of your work, shopping, sleeping, eating, exercising, and digital tracking of communications. There is a need for private appliances and network services, especially for self tracking. And now the patient's digital tracking is done by clinicians, not by patients; the data is about clinical treatment, not day-to-day activity. When you are a patient and a user, if you want to determine the dosage of a drug, which is better? After two weeks of change, you can compare your digital tracking data with the previous two weeks to see how your daily function changes to determine which dose is more appropriate for you. For chronic diseases, depression, memory loss, and Crohn's disease, data on daily activity changes are needed. You are the user of your data; I am the user of my data. My data is for me. Large data is generally obtained from a large n population, while small data is n=me. We need data emancipation, freeing up data for mobile and Web services to yourself. We need an open architecture that produces a wealth of small data on apps and services, just as HTTP standards make www with so many apps and services. Just as mobile apps greatly improve the value of smartphones, personal data-tracking apps should also improve the value of smartphones and the market for small data and personal data warehouses.

It has long been a dream to conquer cancer. Many people now recognize the need to use patient data for personalized cancer treatment. We want to characterize all patients. The DNA of tumor cells causes very different changes in different cancer patients. For example, roughly the same genetic mutation or deletion accounted for only 10% of the patients. Even the same tumor has a different cell mutation. Therefore, the same treatment for many patients cannot be successful. Personalized or hierarchical medications are prescribed for specific patient conditions. Not "the right medicine", but "drug to people." The interactions between genes can cause two of mutations and have a significant effect on the patient's treatment. These individualized therapies need to record and analyze the regularity of individual behavior over time, which is small data.

Of course, the discovery of the general rules of treatment requires large data. Countries in Europe and the United States are planning to compile a database of patient information, not only for cancer treatment, but also for the development of new therapies. Integrating a large number of online databases can boost personalized medication and ease their pain. From large data to the law, with small data to match the individual.

Recently, a lot of blog posts about Big data have been published on the science web, but most of them are general philosophical comments and rarely touch on specific issues. Peng Rombo "The result of the big data is a blessing or a curse?" One article has read 7,816 times, 64 reviews, hit a lot of the key in statistical analysis. Small data are less noticeable. I published a large number of data and small data (131209), by courtesy of the Peng Shron researcher recommended, modified to send to the "China Computer Society newsletter", so far there is no news, the small data may be controversial. The Chinese do not love novelty, may be afraid of small data will weaken the main direction of large data. In fact, "make a different" is the source of innovation. Do not make a different, after others shout, how can innovation? I have always said that large data analysis is not the same as big data technology. Large data technologies are addressed by the IT industry, and large data analysis relies on experts from all walks of life. There is also a saying that small data is a part of large data, small data sets become large data. This is literally to understand large data, small data, technically they are different, the core of the technical problems are completely different. Don't hit a lot of data, put a hat on it, "Big data". Like 20 years ago, systems engineering was also fashionable. which tube is to do the report to talk about a relatively large project, said that it is a system engineering. What about system engineering? " That's hard. " This concept, there is no connotation, what can be explained? So, we should open our minds and study practical problems.