Early detection of cognitive impairment
Dementia may occur associated to some diseases, such as Alzheimer, vascular dementia, dementia with Lewy bodies, Parkinson's disease, etc.
Average costs of dementia in Spain are 19,245 EUR per patient per year, taking into account costs of care, non-health costs and informal care.
The most prevalent disease in this area is Alzheimer's. It rises 50% of people with dementia:
Cognitive decline appears before the development of dementia.
User device (PC, smartphone, …)
Wireless access point
Collects and analyzes data in real time
ERIS Innovation proposes cognitive assessment by using ICT.
Patient uses a connected device:
Using this solution, we may make an early diagnoses of cognitive decline, improving the quality of life of patients.
It would be a very feasible solution oriented to home for the elderly, both public and private. Also, it could be used at community health center.
Block Detection associated with Parkinson’s Disease
According to World Health Organization:
In the case of Parkinson's Disease (PD), falls are because of the people instability.
People with PD suffer Freezing of Gait (FoG) since first motoric symptoms occur. These episodes are brief and intermittent, and get rid of walking.
FoG appears at the start of the gait or changes in it (changing the speed, direction, walking in tight spaces, etc.).
FoG are the greatest loss of autonomy of people with PD, since in many cases, it results in a fall.
Wireless Access point
Collects and analyzes data in real time
At ERIS Innovation we propose the use of wearables for detecting patient movements.
These movements will be sent through a Wireless way to a data sink, which will pre-process them for defining an ideal set of characteristics.
These characteristics will be used for implementing a classification algorithm thought machine learning techniques
This algorithm will detect the beginning of the FoG in advance. It is stored in data sink.
Big Data Analytics for Air Traffic Control
According to International Air Transport Association (IATA):
A key aspect of this system is Air Traffic Control (ATC), vital for safely directing and navigating airplanes through the local airspace, during take-off and landing. ATC applies separation rules to the aircraft that they direct. Separation rules are used to regulate the distance between other airplanes and that aircraft by requiring a minimum distance among them. This is for increasing safety and reducing unnecessary risks for pilots and passengers.
To solve the saturation of the air transport system, two initiatives are in place to completely overhaul airspace and its air traffic management (ATM): Single European Sky ATM Research (SESAR) in Europe and NextGen in USA.
SESAR's target concept relies on a number of new key features:
A whole set of new tools are needed to understand, model, plan, forecast and control the air operations under the new paradigms.
At ERIS Innovation we apply traditional statistics in addition to Machine Learning techniques such as; Bayesian classification, cluster analysis or multivariate regression, employing innovative methods like Deep Neural Networks or Recursive Neural Networks to recognize patterns, classify or detect anomalies in flight data.
These data analytics tools are combined with powerful and realistic simulations of the aircraft dynamic behavior to obtain useful information like operation security margins, fuel consumption and flight efficiency.
This rich set of analysis and simulation tools allow the development of several solutions:
Online Fraud Detection
The use of online payment mode such as online banking, debit card, credit card etc. has become hugely popular and important in day to day shopping and e-commerce activities. On the other hand, online fraud has become a serious issue in financial crime management for all online businesses.
The companies and financial institutions are losing huge amounts due to fraud while fraudsters continuously try to find new rules and tactics to commit illegal actions.
According to different studies, annual fraud costs approximately for U.S. retailers reached $32 billion in 2014. Retailers lost an estimated 1.3% of revenue in 2015, more than double the rate of 2014.
It also proved that up to 25% of declined sales transactions for e-commerce merchants were actually good sales to start, indicating that the most of current fraud detection systems have huge rates of false positives
The important fact is that the fraud is an adaptive crime, so it needs special methods of adaptive intelligent data analysis to detect and prevent it continually.
At ERIS Innovation we have considered several supervised and unsupervised approaches toward implementation of novel AI based anomaly detection and pattern recognition methodologies for developing cutting-edge online fraud detection systems.
hanks to the hybrid structure of these solutions which gets benefit from the learning capabilities of Artificial Neural Networks (ANNs) and reasoning abilities of Fuzzy Logic based models (FL), these systems are able to learn from cybersecurity experts and fraud attacks in the past and then making an updated bank of recognized fraud patterns as well as being able to constantly monitor and notify the suspicious online activities.
Short Time Electricity Prices and Demand Forecasting
Electricity is a very special commodity, practically non-storable. The power system stability requires a constant balance between production and consumption which means that the demand must be satisfied continuously.
The process of deregulation and the introduction of competitive markets have reshaped the traditionally monopolistic and government-controlled power sectors. Nowadays, in many countries all over the world, the production and sale of electricity is traded under competitive rules in free markets.
If producers and consumers are able to make reliable predictions of electricity price, they can develop their bidding strategies and their own production or consumption schedules in such a way to reduce the risks or maximize the profits.
The costs of over/under contracting and then selling/buying power in the real-time balancing market are typically so high that they can lead to huge financial losses or even bankruptcy.
Electric utilities are the most vulnerable, since they generally cannot pass their costs on to the retail consumers. Consequently, prediction of electricity demand and price are significant problems in this sector.
At ERIS Innovation we have tried out several methods for short time electricity prices and demand forecasting, from hard computing techniques including time series models and dynamic regression to the soft computing approaches, sometimes referred to as the artificial intelligent tools such as fuzzy neural networks (FNNs) and in general hybrid intelligent systems (HIS).
Our models have been tested on the electricity market of Iberian Peninsula, mainly Spanish market, which is commonly used as the test case in several electricity price forecasting studies
The obtained results have shown great statistical characteristics as well as high accuracy for a day-ahead hourly price forecasting.
Datacenter Power Usage Energy Optimization
With the growing popularity of new Cloud applications and services, IT networks and data center's computational demand have experienced a rapid increase, playing the central role in business opportunities and digital services.
Data centers represent a critical pillar in businesses for companies in a wide range of industries, raising more and more everyday their economic impact in business operations.
Energy consumption of data center facilities has reached the 2% of the global outcome and this energy is mainly produced by non-renewable sources, having not only a high repercussion in the economy but also in the carbon footprint and the thermal impact.
The main sources of energy consumption in data centers are due to both computational and cooling contributions.
Hybrid Multi-Stage Adaptive Scoring
Scoring models have been used to define the risk and loyalty level of customers, consumers and new product buyers and specifically in the financial sector, credit scoring is a method of evaluating the credit risk of loan or credit applicants.
Nowadays, credit scoring plays a vital role in economic growth by improving easy and fast access to credit markets, lowering the price of credit and reducing the economic risks.
Credit scoring has gained more and more attention as the credit industry can benefit from improving cash flow, insuring credit collections and reducing possible risks.
Traditional scoring methods require that a manual variable reduction/creation process based on trial and error is repeated for each update after model development.
Moreover, it is difficult to develop an adaptive credit scoring model based on the current linear methods due to their inherent constraints. Therefore, a dynamic modelling process using state-of-the-art hybrid techniques could be useful for the scoring tasks.
At ERIS Innovation, firstly we used a multi-objective optimal variable selection in order to obtain the best subset of features with the minimum number of variables in that subset. Secondly we have developed a Hybrid Multi-Stage Adaptive Credit Scoring model applying some advanced metaheuristic classifiers.
Optimization of The Distribution Networks
Over the past decades, many businesses have struggled with the number of fundamental issues such as sourcing, production, warehousing, distribution and customer service. However, in today's global demand-driven market, Supply Chain Management has become an urgent need for product and service providers all around the world.
The increasing customer expectations, highly competitive global marketplace and rising transportation costs due to inflated and volatile fuel prices are driving the development and investment of new distribution network designs.
In the case of one US-based manufacturer of industrial materials, a supply chain management study resulted an annual warehousing costs reduction by about $300,000 for an investment of approximately $100,000 in the network modeling and warehouse redesign. Because of the global nature of the company, there were also significant potentials for additional transportation savings, approximately $2.5 million annually only by investing about $420,000 to develop and implement optimal processes.
Distribution Network Optimization guarantees the quality and safety of the product flow from the manufacturer to the retailers and end users. However, Supply Chain and Distribution Network Optimization are possible if they rely on an accurate forecasting of the needed products or services. Since the prediction of the product demand and supply is an input for the optimization models, the forecasting methods are extremely important to yield highly reliable results.
At ERIS Innovation we have developed a comprehensive solution for the product flow management which contains a forecasting module, an optimizer unit as its central component, as well as a risk analyzer section to monitor the certainty and uncertainty level of the obtained optimal solutions
There is an important need to create a structured way to collect, process, and analyze a set of data to make better performance and increase functional capabilities of the facility.
Modern buildings are full of data collection systems, from building management systems -BEMs- which captures temperature and humidity levels, to access controllers, which collect occupancy statistics, to other measurements too.
Big data describes large, complex sets of data that if interpreted properly, can provide insight into how a building is performing from an energy standpoint float-right2. Building owners are realizing that the metrics they use to evaluate the assets in their portfolios are changing.
Shifting customer demands, climate change mitigation and sustainability goals, power reliability and resiliency concerns, and budget constraints are driving demand for intelligent building solutions.
Therefore, there is an important need to create a structured way to collect, process, and analyze a set of data to make better performance and increase functional capabilities of the facility.
At ERIS Innovation, we are developing Big Data analytical solutions to the numerous problems for the buildings. Decision making becomes easier if we have a set of data collected through surveys, smart meters, IT networks, and occupant behavior.
The main emerging added value solutions for smart buildings are included in our solution: