These days everybody is talking about AI...it seems the only logical choice for any industry. So, why even to bother with statistics when there are much fancier machine learning/ neural network applications available?
Well, it depends what is the task at hand and which kind of data is available!
I guess, anybody prefers to (sometimes) riding a Porsche. But when the task is plowing a cornfield, a Porsche might not be the best choice! The same is true for statistical models and machine learning/ neural network applications.
Commercial real estate industry is one of the largest industry sectors, but when it comes to the empirical data generation, also one of the "slower" meaning that not too many data points are being produced as it is the case, for example, in the residential part. This makes simple application of AI technologies such as machine learning impossible, without using other methods, such as statistical modelling.
On the other hand, other kinds of available data sources, currently not being used in the commercial real estate investment management are ideal for the AI tools application. This makes combining statistical modelling with AI the optimal solution for commercial real estate investment industry,
Statistical Modelling vs. Machine Learning
Here is a comprehensive summary (references below) which describes the difference between those two worlds:
Statistical Model is the use of statistics to build a representation of the data and then conduct analysis to infer any relationships between variables or discover insights. It is a method of mathematically approximating the world. A statistical model will have e.g. sampling, probability spaces, assumptions and diagnostics to make inferences.
Machine Learning is the use of mathematical and/ or statistical models to obtain a general understanding of the data in order to be able to make classifications and predictions.
So the crucial element of distinction is the data available and suitable for both of the approaches. Statistical models get a grip also on small data sets, while Machine Learning/ Neural Network Applications need a huge amounts of data in order to be able to identify the right patterns and to make sense of this learning effort in terms of, e.g. classification, forecasting, synthetic image generation and similar.
Commercial Real Estate Applications
When we talk about digital technical twins in the commercial real estate business, there we have an enormous potential within the framework of machine learning.
Sensors implemented in all parts of a building deliver millions of data points within a short period of time. These amounts of data are incomprehensible for a human brain to make sense out of it, but ideal for machine learning/ neural network methods, which can learn on the data flow and distill hidden patterns in no time.
But if we move to the other end of the spectrum, investment management in commercial real estate is a part of the industry where, due to low data frequency flow this kind of simplified approach is not possible. Here we have to use statistical modelling tools.
On the other hand, also in this segment combining AI with other tech tools offers huge opportunities to analyse the market in ways which are not possible by human action. Namely, there are lots of market parameters which are impacting commercial real estate investments, currently not taken into account by the market reports. These elements are an ideal data stream for AI driven advanced analytics.
The final answer is in the right combination of statistical modelling (solving the low data frequency challenge) and AI (using the abundance of overall market parameters data points) in order to get deep market insights, objective future developments predictions and uncover hidden market connections.
While AI is something that is being talked about a lot lately, statistical modelling is not, although it is very often a necessary tool for low data frequency industries, such as commercial real estate. Statistical models are ideal to offer very reliable insights into future market developments and help to reduce risk exposures. This, in combination with AI, this approach gives the investors an opportunity to take appropriate market positions in due time before market developments actually happen.
So, implementing these advanced solutions, e.g. in the form of predictive analytics, in the workflow of real estate investment management is the natural next step. It is a crucial support system for decision makers running commercial real estate investments.
The alternative is insufficient and outdated, although sadly still used by a large portion of commercial real estate investors: waiting for very rare empirical market reports and based on them, subjective guessing about future developments. This is more or less of speculative approach, while money, i.e. investor’s capital, has to be committed and face risks in the mid- to long run.
A Word of Caution
Although the benefits of statistical models incorporated in the investment management of commercial real estate investors are obvious, there is a need for various considerations when using those tools so they are suitable for the respective task and appropriate for the given data.
To give an idea of some of the challenges in statistical modelling, here are three examples:
State-of-the-art ARIMA-GARCH models need at least 800–1,000 data points in a time line. This works excellent with share price quotations but in general is out of scope for real estate market data.
The incorporation of dependency structures among market parameters could fail when the ratio market parameters/ observations is too stretched. In this case, the structure might fail or deliver biased results.
Trends in time lines not treated properly while developing the statistical model could lead to biased and therefore wrong results.
So, when using statistical modelling the right amount of industry expertise is necessary in order to find the right methods and to be able to derive accurate results. Only custom made solutions in combination with AI can be a basis for offering commercial real estate investors the right investment frame which will enable them to actually see through the market and detect upcoming risks and opportunities.
Statistical Modelling vs. Machine Learning by Asel Mendis, KD Nuggets/ August 14, 2019