Information Analytics

Octillion provides analytics solutions to improve your decision making by an order of magnitude. Your decisions are not based on gut-feel anymore. Rather, you base your decisions on precise and real-time information which substantially reduces the risk of going wrong. You never have to accept any decision that is less than perfect. Our solutions reduce the cost of doing business and improve the quality of your products and services to your customers.

Octillion solutions are unique because:

Expertise - Octillion executives and engineers are highly experienced professionals
Agnostic Approach - Octillion is technology-agnostic; therefore, solutions are built with the single purpose of delivering the desired services
Standards-based development - This approach facilitates easy-to-support and upgrade technology
Flexibility - Octillion provides knowledge transfer to meet your needs
Industry expertise - Octillion has expertise in financial services, healthcare, manufacturing, and governments

Finance Industry Solutions:

Treasury and Funds Management

This solution sets up message format and information work flow systems in both fully automated and semi automated environments.

Solution develops a dynamic investment management and dynamic debt management frameworks using time series analyses and different optimization models, and;
Constructs an auction like framework which captures the knowledge of the managers on the floor to capital allocation

Anti-Money Laundering through Pattern Recognition

The solution enhances the efficiency of intervention by building models that can reliably flag/ highlight transactions related to laundering.

After modeling “business as usual” we analyze importer/exporter historical transactions using soft computing techniques (fuzzy as well as statistical techniques)
Looks for systematic deviation (not just individual outliers)
Allows manual addition to the patterns and let the system also learn from manual inputs.
Uses mining techniques, such as, temporal mining methods, radial-density derivatives, statistical methods, unsupervised clustering and supervised learning methods

Determining the value of derivatives like instruments

Solution assess the value of the instruments (Bonds, Money Market Instruments, and Bills) offered by its affiliated financial Institutions as security for over draft or other purposes. These instruments work like derivatives in the sense that the overdraft limits are determined based on their value.

The solution determines instrument value and identifies the duration for which it should remain with central institution
It uses combination of VAR, LP and NLP approaches for optimizing risk maximizing returns and determining optimum duration profile

VAR is used for value determination and risk tracking
LP is used allocate sector wise or duration wise allocation
NLP is used to build up optimum portfolio

Default Risk Assessment

This solution improves the existing rule based credit policy approval score card so as to control default and enhance approval rates

Solution integrates data from customer application with customer demographics, product detail and delinquency information.
Builds a model to discover underlying patterns of customer behavior and identify key factors contributing to high delinquency rate.
Predictive models were built to generate future default that would map to the application information.
Sophisticated reject inference techniques were used to get around the problem

Increasing Debt Collection effectiveness using Decision Trees:

To increase debt collection on portfolio in default, while using minimum collection resources.

Decision tree model was used to study the rule based classification to segment the delinquent customers in the following classification

Early delinquency buckets.
Higher delinquency buckets.

Correlate debtors’ behavior with transactional history to segment the portfolio of debtors.
Use the segmentation of debtors to predict the behavior in near future.

The Decision Tree model used to segment the potential self cures and those which had high potential to roll over to high delinquency buckets.
Provide key indicator variables used for segmentation and scoring model. Variables like maximum past delinquency, credit limit utilization, average pay down ratio, balance outstanding etc to develop collection model.

A decision support system for treasury optimization

This solution maximizes investment return & minimizes borrowing cost in corporate treasuries, subject to treasury policy constraints (e.g. only 20% in BB+ or lower, 90-day max average weighted maturity), and conditional on cash flow forecast

Applicable to bank’s internal book, money market funds, corporate treasuries, etc.
Exploits a combination of optimization strategies, from LP to A* to simulated annealing

Risk Assessment in Insurance

Insurance carriers currently use historical information and demographic information to determine risk. They also use trend projections and conventional approaches for each area in isolation. These methods are susceptible to natural dynamics, global politics and other factors beyond their control.

Octillion solution proposes to use:

Demographic profiles and patterns of future customers in the next 15 years
Knowledgeable consumers with wide variety of behavioral patterns, expectations and technology sophistication (Boomers, Gen X, Gen Y etc)
Use of Dynamic products – JIT to long term policies
Active Risk Management and Mitigation using extended forecasting models spanning multiple areas
Leverage emerging and futuristic technologies and forecasting models for pre-emptive risk mitigation

Retail Industry Solutions:

Decision support solution

To develop a warehouse decision support for interaction between the shop floor managers and decision makers and develop analytics models for predicting the procurement and cash flow gaps.

Architecting and setting up a web 2.0 framework that enables communication, transactions in addition to generating useful secondary data that helps the system make valuable predictions about outlet requirements
Designing a centralized data warehouse and generating informative reports
Building a unique predictive tool combining time series analyses, supervised learning and floor manager inputs for better decision support

Ecommerce System

Build an Ecommerce system embedded with data and image mining for improving user responsiveness to online catalogs.

Interpreting clicks, mouse movements and time spent as relating to user interest.
Learning characteristics for the subset of interested items, based on attributes like color, size, type & image of item etc.
Applying the model to assign an interest score to every item in the catalog and rank items while displaying based on this score.
Building a separate model for each user.
Adjusting models over time, as users spend more time on the catalog.
Where sparse data is available for a specific user, extrapolating interest patterns from users with similar demographic profiles.

Enterprise Web 2.0 Solutions:

Capabilities and domain expertise

Business Process Automation

Customer Relationship Management, Supply Chain Management, Sales Force Automation & Workflow Management

Content Management Systems

Web Content Management, Enterprise Information Portals, Document Management, Course Management & File Management

Social Networking

Online Communities, Blogs, Internet Forums, Online Chat, Matchmaking Portals & Wikis

B2B / B2C Portals

Oracle Portal, SharePoint Server, Sun Portal, Customized Portals, e-Market Places, Online Magazines, Recruitment Portals, Entertainment Portals & e-Commerce Portals

Knowledge Management

Information Management Systems, Help Desks, Corporate Intranets and Extranets & Data Repositories

Communication & Digital Media Distribution

Web Conferencing, Instant Messaging, Audio/Video Streaming and Broadcasting & Mailing Systems

Trade Portal

Complete Market Place to buy or sell physical or financial products and packages including search and auction engines
Financial Instruments Auction of Purchase Order (PO) financing, Invoice (INV) factoring, discounting and forfeiting financing, Letter of Credit (LC), Loan and Collection financing
Trade Service Providers listing including links and current stock market data
Complete My Business directory including trade partners, history and status of trade and financial components, statements, reports
Single Sign-on and Administration, secure access to sensitive information

Warehouse Planning Software

Multi-depot and multi-site operations
Subcontractor Management
Transport and Route Planning
Carrier Selection
Order Management
Vehicle and Driver Utilization
Mobile Solutions and SMS
KPIs and Business Intelligence Reporting
Dashboards and Automatic Alerts
Visibility and Collaboration Systems Integration

Text Mining

Objective: To build models to (a) cluster documents of like content, and (b) classify new documents into these content types automatically
Methodology: Hidden Markov models, statistical feature extraction, text preprocessing, principal component analysis, rule extraction & decision trees
Benefit: Speed and ability to handle thousands of documents with millions of words, ability to utilize human knowledge as part of the model

Other services using Statistical Analytics and BI technology

Loyalty Analytics
Credit Scoring
Debit Scoring
Portfolio Loss Forecasting
Cross Selling Modeling
Buyer behavior in selecting a financial instruments
Economic Analysis
Estimating the value of securities and bonds
Analyze risk associated with loans
Analyze risk associated with various investments
Response based market segmentation
Optimum allocation of products per store among a chain of stores
Forecast the sales of various products of a specific brand in a store
Brand scorecards
Price elasticity and pricing strategies
Product preference analysis
Positioning the consumer products in a space based on similarity ratings
Position products and individual ideal points based on product rankings or ratings
Products purchased positioned in a space with different customer groups