Join the Community

21,017
Expert opinions
43,852
Total members
328
New members (last 30 days)
111
New opinions (last 30 days)
28,287
Total comments

MDM in the Cognitive world

Be the first to comment

Staying relevant in today’s times necessitates digital enterprises bridge the gap between data generated internally vs externally whilst continuing to drawing meaningful insights. As copious amounts of data continue to get generated by the second and one cannot undermine the significance that MDM brings to the table in the age of cognitive computing

As much as most enterprises across the world have embarked on their MDM journey, the maturity of the MDM program and hence the ability to harness the value of the diverse data is what separates the chaff from the grain. Sample a retailer, where they would want to conduct targeted campaigns across specific customer segments, Under these circumstances until the digital footprints of customer interactions, and customer intelligence across multiple channels, touch points, as well as social networks are not correctly captured the campaigns will not necessarily yield the desired results. The basis of all which is the core Customer dimension or master data as we call it  

Per recent statistics -‘70% of a Data Scientist’s time is spent on data collection and preparation, rather than building and deploying predictive models’   

To stay true to its spirit, and to reap the many benefits of cognitive computing, lies the underlying rigor in managing the master data and this includes the elements of syntax, semantics, discovery, integration and quality

Semantics and taxonomy

Modern data modelling techniques along with ML  help translate schema from disparate sources into holistic data representations applicable to multiple types and origins of data used to form a golden record of Master data . The use of Graph technology has further brought about a paradigm shift in the data modelling space by going schema-less and driving multi-dimensional views that are vital to all use cases for the ‘System of record’

Data Architecture 

Incorporation of AI and ML models implies the reimagination of the storage and consumption of the core data. Machine Learning systems requiring a feed from several siloed stores of master data, reeks of under-efficient and error prone processes. Add to that legacy systems and inconsistent definitions spell bottlenecks as far as performance and response times are concerned 

Efficient and accurate processes demand a single system of record that can be referenced.

Quality of Data

Cognitive typically relies on repeatable and correlative based methods, requiring clean , accurate  standardized data, In the absence of a holistic MDM system not only are organizations dealing with incomplete and inconsistent data but also heavily compromising on the time to value ratio despite investing heavily in cutting edge technologies

As much as  MDM is responsible for maintaining that standard data catalog within the enterprise, there is great deal of work that can be potentially offloaded to AI and ML technologies.ML techniques can help identify  probabilistic matches of multiple data records, which in turn helps in quicker analysis by determining the  system of record

Data Stewardship

Steward-ship activity is one of those unavoidable but compulsory pursuits that is required to maintain the sanctity of master data. But with time and mounting volumes, this can take a toll on overall throughput and efficiency And that’s precisely where ML can come to rescue by prioritizing, routing jobs to Steward groups based on the interpretations from previous results

Extended Data

With a multitude of transaction, social media feed, big data, MDM in itself has immense potential to be expanded to build a 360 degree view, the imperatives for which lie in AI and ML.

AI and ML demands an expansive field of opportunity as far as data is concerned and of which MDM plays a very significant role, since cognitive as often mis-understood is not  as much about voluminous data alone but about unearthing data such that the valuable portions are leveraged by the learning datasets

Simply put MDM is that common thread within the data fabric which elevates the use of the power of information to help organization gain a competitive advantage

 

External

This content is provided by an external author without editing by Finextra. It expresses the views and opinions of the author.

Join the Community

21,017
Expert opinions
43,852
Total members
328
New members (last 30 days)
111
New opinions (last 30 days)
28,287
Total comments