Automated Reading in the Big Data Era
In this information explosion era, enterprises need to collect and analyse various types of information, e.g. customer feedback regularly, and manage massive internal data, e.g. archives. These data come in various formats and structures, often written in a mix of languages including English, Traditional Chinese and Simplified Chinese.
Handling unstructured text and managing archives efficiently are becoming big challenges to enterprises. Traditionally, huge amounts of manpower are necessary to handle these problems, leading to very high operating cost.
ClusterTech has developed PRISMA (A.I. Reading Robot) – an Artificial Intelligence technology for text processing, which revolutionises text processing and cost-effectively solves the above problems, enabling organisations to handle their data systematically, increase operating efficiency and reduce time and labor costs.
3 Core Technologies of PRISMA
PRISMA is an innovative and state-of-the-art text processing technology. It consists of 3 core technologies including “Search Engine”, “Text Mining”, and “(Trilingual) Natural Language Processing (NLP)*”. In the past, text documents usually relied heavily on manual reading and handling. Now, structured and unstructured text data including Cantonese slang, idioms and buzzwords can be understood and automatically categorised by PRISMA on the fly.
*The Natural Language Processing (N.L.P.) Technology in PRISMA is unique. It is capable of processing trilingual texts (Traditional Chinese, Simple Chinese and English) simultaneously, and is tailor-made for the multi-cultural Hong Kong market.
PRISMA Key Functions
- Build a relevant taxonomy (text classification dictionary) based on industrial features
- Read text data, recognise key words and context fast
- Detect and correct typos automatically
- Understand structured and unstructured text data
- Classify key contents to proper categories with an accuracy exceeding 80%
- Detect sentiments objectively with an accuracy
- Tuneable for newly added keywords, customised typos and alias
PRISMA-Powered Text Solutions
Building on PRISMA, ClusterTech can customise text processing solutions according to enterprise requirements. PRISMA-Powered Text Solutions consist of 2 major elements : “PRISMA-Powered Software” and “ClusterTech Professional Analytics and Consulting Service”.
Features of PRISMA-Powered Software
- Graphical user interfaces to display analytic results with different types of charts and formats
- User-friendly design suitable for users with or without technical backgrounds
- Software and artificial intelligence upgrade
- Scaleable data storage design to accommodate big data
- Available on premise or as a service
- Highly compatible and customisable for third-party software integration
Features of ClusterTech Professional Analytics
and Consulting Service
- The ClusterTech professional service team, comprising experienced software engineers, data scientists and business consultants, work closely with enterprise users to review their business needs, e.g. key performance indicators (KPI), evaluate keywords and categorisation alternatives, develop suitable taxonomies and then tune PRISMA to achieve effective and accurate classification of text data.
- Based on the classification results generated by PRISMA, our professional team can also conduct further analysis by data mining, mathematical modelling, etc, to identify hidden data patterns, reveal correlations in text data, and forecast trends, arriving at actionable insights to enterprises.
- With the use of big data technologies, our system integration services help enterprises consolidate customer data from various sources in one place, automating data exchange across systems and allowing users to make the best use of data.
Managing and understanding text data with Artificial Intelligence (A.I.) :
- Increase workforce productivity and reduce labor costs
- Achieve real-time performance and improve user experience
- Enable enterprises to make timely strategic decisions based on objective and
accurate classification of text content
Applications of PRISMA-Powered Text Solutions
Case 1 :
Social Listening – Understanding massive amounts of customer feedback
efficiently and accurately
PRISMA, while screening customer feedback, uncovers the correlations between text data and enterprise objectives by extracting and categorising key information. Leveraging PRISMA, enterprises can gain a clearer and more comprehensive insight into customers and their needs from a large volume of text data efficiently and accurately. As a result, enterprises can make fast and timely business strategies, improving customer experience as well as achieving long-term success.
Case 2 :
Chatbot – Providing intelligent and real-time online chatbot
PRISMA can automatically respond to customer enquiries in text based on the pre-set service categories. It can be widely applied on e-services platforms, e.g. online customer service platforms. Enterprises may not only reduce their labor cost, but also provide high service quality during non-office hours.
Case 3 :
Knowledge Management – Managing archives intelligently
PRISMA detects and classifies keywords into proper categories during reading documents. It can get rid of biases on text categorisation and shorten the time of organising documents when compared to manual processing. As a result, enterprises can centralise documents and build systematic archives to facilitate staff to search relevant information by keywords.
Interfaces of PRISMA-Powered Software
- User Interface is simple and easy to use.
- Graphical user interface displays analytic results in various types of charts.