USE CASE
Auto-classification
“It’s a given that beings are inherently inconsistent. Auto-classification can help to reduce the impact of human inconsistency on information management, resulting in more accurate, consistent, and efficient processes.”
Why you need auto-classification
Auto-classification is the process of automatically categorizing or classifying data or documents based on predefined criteria. Automating the classification process will bring significant advantages to your organizations, including:
- Increased efficiency and productivity
- Improved coverage
- Greater accuracy
- Enhanced findability
- Easier records management and data protection
- Significant cost savings
How Synercon can help
Synercon have substantial experience in deploying auto-classification to deliver records management outcomes, chiefly in the SharePoint environment. We can help you generate high quality fit-for-purpose metadata which will improve findability, automate appraisal and disposal, enable migration and apply information protection.
Auto-classification Tools
We work with a suite of tools to deliver our auto-classification solutions.
a.k.a. software by Synercon simplifies the task of developing taxonomies, thesauri and ontologies for deployment into an auto-classification engine and into the SharePoint Term Store. With a.k.a. we codify your classification schemes and synonyms into machine readable formats.
Synercon SuperSaurus enables us to fast track the production of taxonomies, thesauri and ontologies for auto-classification.
DiscoveryOne Content Enrichment engine for processing auto-classification across multiple environments and tagging records within the SharePoint.
Auto-classification Techniques
Synercon have developed considerable know-how for deploying auto-classification into the Office 365 and SharePoint environment.
Our auto-classification methodology is based on a faceted classification model from which we generate fit-for-purpose metadata, which enhances findability and feeds records management and data protection processing.
We employ a deterministic rules-based approach to auto-classification which is upfront, observable and measurable.
We have developed simple-to-use techniques for analyzing and measuring coverage and tag accuracy.
We integrate with SharePoint’s Term Store and Managed Metadata to truly enable the SharePoint search experience.
We advocate an iterative, measured approach to deployment, so that benefits can be realized from the beginning.
Auto-classification Training
Auto-classification is a complex process that requires a combination of domain expertise in taxonomy development and information governance, along with of knowledge of rule building, data architecture and information systems.
Synercon delivers integrative learning – education and training that connects the multiple branches of information governance.
Our training programs are designed to facilitate knowledge transfer between disciplines. By incorporating the strengths and lessons learned of each profession, we help you build information governance instruments that are fit for today’s digital environment.
Synercon Training Courses
Frequently asked questions
What is auto-classification?
Auto-classification is a methodology that that classifies content by analyzing its text and other properties, and through the application of rules and/or algorithms, automatically assigns labels (or tags) to the document based on a pre-defined class.
Auto-classification can use either machine learning models, where the system defines its own set of rules that are based on data outputs, or rule-based models, which produce pre-defined outcomes that are based on rules coded by humans.
Each system has its own advantages and disadvantages.
Machine learning models are probabilistic and use statistical rules rather than the deterministic approach of a rule-based AI model. Machine learning systems require significantly more data than rule-based models to achieve accurate results.
Rule based models are recommended when there is a high chance of error and when fast results are required. The Synercon solution uses rule based models which deliver the high accuracy required for effective records management.
What are the benefits of auto-classification?
Automating the classification process bring significant advantages for organizations, including:
Increased efficiency and productivity
Auto-classification enables business users to get on with their day jobs because they just don’t have time to classify the high volumes of content that they manage.
Even if you gave the job to an export, it would take them 2-3 years to process what an autoclassification engine can process in one night.
It’s much more efficient to have your classification experts, build the taxonomies and the rules, than to manually classify content
Improved coverage
Auto-classification will vastly improve coverage by ensuring that ALL documents are classified, regardless of the volume or complexity of the documents. No documents in your crawl will be missed or overlooked.
This can help ensure that all relevant documents are included in the records management system, reducing the risk of missing important documents or information.
Greater consistency and accuracy
Auto-classification delivers consistent metadata, because the rules that are established for auto-classification are applied consistently across every piece of content.
Accuracy and precision are improved by measuring results and refining and optimizing the classification rules and criteria. The more knowledge your feed into your engine, the better it will classify.
Synercon’s solution regularly achieves accurancy in the range of 80-95% for our generated metadata.
Enhanced visibility and findability
Synercon’s solution generates high quality fit-for-purpose managed metadata into the SharePoint environment. This means that you can effectively use all of the enhanced SharePoint search experience such as Views, Filters and Refiners.
Enables easier records management and data protection
Auto-classification enables records management processes when records are tagged with high quality fit-for-purpose metadata, which provides information about the content, context and significance of records. It is this metadata which provides the foundation for aut0-appraisal of records..
Significant cost savings
Auto-classification offers significant cost savings in data and records management, particularly in terms of reducing the time and resources required for manual classification, increasing productivity, reducing errors and rework, reducing storage costs, and improving compliance.
How accurate is autoclassification ?
The accuracy of auto-classification depends on a number of factors, including the breadth and quality of the content, the complexity of classification criteria or rules, and the effectiveness of the auto-appraisal algorithm or system. There are many techniques available to choose from.
Synercon have evolved techniques for analyzing data sets and building quality rules that eliminate ambiguity. With our techniques we can achieve accuracy of over 80 — 95% for text-based content.