HOME /

Financial Forecasting using Machine Learning

 

Financial Forecasting using Machine Learning

 

What is ML:

Machine Learning (ML) is a tool to extract knowledge/pattern from data. We can use ML for financial forecasting, to predict supply/demand/inventory of the market, and improve business performance.

How ML create value:

ML can analyze historical data to understand the demand, supply, and inventory, then forecasts the future's demand, supply, and inventory. ML can forecast client's budget and several other economics’ indicators, thus help the business improving their performance.

ML Past & Future:

According to Gartner, ML has just reached its peak of hype cycle, and will enter plateau in 2 to 5 years. ML is a subset of AI, with the developing of GPU, ML is evolving into deep learning with faster speed, better performance, and lower cost. There are three subdisciplines of ML: supervised learning, unsupervised learning, and reinforcement learning.

ML1
How to implement it?

There are several algorithms available for ML forecasting, some of the most popular are Multi-Layer Perception (MLP), Time Series Forecasting, Window Method, Gaussian Process.

Forecasting Processing Using MLP:

1) Create the MLP network.
2) Training the MLP Network.
3) Testing the MLP network.
4) Generate the prediction.

ML2

Currently available technology for ML:

Many big vendors are offering several options and technologies for ML at a reasonable cost for small companies and big companies, range from open-source-tool to enterprise-Cloud-platform.

ML as cloud platform as service (Paas): Microsoft Azure, Amazon AWS, Google Cloud, IBM Bluemix.
ML as service (SaaS): Dato, rapidminer, predictionIO.
Deep Learning with GPU: Microsoft Azure, ErsatzLabs.
ML for Investment Finance: Bloomberg, Quantopian, Dataminr, Kensho.
ML Open source Tools: TensorFlow, SparkMLlib, Microsoft DMTK, SciKit.
Top 10 forecasting solution according to CFOTechOutLook: 9Dots, Adaptive Insights, Arcplan, Axiom EPM, Centage, MRI Software, OneStream Software, Solver, Tidemark Systems, Xlerant.

MLLandScape
Recommendation:

Consider all the benefit and technology trend; Company should invest in a machine learning team and develop a machine learning model to improve their business's performance.

-A small company should start out using ML Software as service (SaaS) or Platform as service (PaaS) as it expected a small upfront investment.

 -A big company should invest in ML as Platform as service (PaaS) as it expected a   bigger upfront investment.

-Big company that adopted an aggressive growth strategy should develop a talent team, build their deep learning enterprise system by GPU or PaaS, to create more competitive advantage.

 

Reference:

1. Top 10 Budgeting and Forecasting solution providers - 2016, 2016, CFOTechOutlook. Retrieved from: https://budgeting-and-forecasting.cfotechoutlook.com/vendors/budgeting-and-forecasting-solution-providers ­2016.html

2. Hype Cycle for Emerging Technologies, 2016. Retrieved from: https://www.gartner.com. Gartner ID: G00299893

3. Machine Learning in Financial Forecasting, HainDrich Henrietta and Vezer Evelin. Retrieved from: http://www.cs.ubbcluj.ro/~csatol/mach_learn/bemutato/HaindrichRist_FinancialForecasting.pdf

4. What's the Difference Between Artificial Intelligence, Machine Learning and Deep Learning, 2016, Michael CopeLand. Retrieved from: https://blogs.nvidia.com/blog/2016/07/29/whats-difference-artificial-intelligence ­machine-learning-deep-learning-ai/

5. An executive’s guide to machine learning, 2015, Dorian Pyle and Cristina San Jose.

6. Which PassS is Wining the Machine Learning and Artificial Intelligence Race, 2016, Jesus Rodriguez, Retrieved from: https://medium.com/@jrodthoughts/which-paas-is-winning-the-machine-learning-and-artificial ­intelligence-race-2640e1e96eed