Data Mining to measure the financial crisis in Governments

When it comes to evaluate the financial condition of the government, the process is focused in 4 dimension which are

• Sustainability
• Solvency
• Flexibility
• Financial Independence

For the data mining of the crisis, use of rule based induction algorithm called CHAID is famous. This algorithm provides the efficient data mining technique for segmentation or tree growing.For the levels of budgetary sustainability the most significant variables are related to the current margin, together with the importance of capital expenditure in the budgetary structure.The short-term solvency directly depends on the liquid funds possessed by the entity body. The flexibility depends mainly on the financial load per inhabitant of the municipality, on the total sum of fixed charges. Finally, financial independence depends fundamentally on the transfers that the entity receives and on the fiscal pressure, among other elements.

The detection and rectification of financial crises in local authorities is found to be fundamental interest for public-sector managers. To decide if a local authority is well managed or ill managed, it seems to be necessary to take into account a series of external factors that are dependable in this respect. In general, a control system of these characteristics makes it possible to advise different types of users of the existence of financial tensions. And users may be interested in  including public-sector managers in authorities responsible for supervising the financial situation of town and city councils, or senior officers, who need to know how resources are being managed, and how this is done in comparable councils. In order to determine whether a local authority is experiencing a financial crisis, we consider the concept of financial condition, which is measured by means of different elements, listed out  above. This concept can be divided into other, more specific, aspects, such as those of flexibility, sustainability and vulnerability (Greenberg and Hillier 1995; CICA 1997). Finally, long-term solvency is measured through the incorporation of a considerable period of time into the indicators considered. However, there is a problem concerning the measurement of the elements that constitute financial condition, namely the non-existence of an instrument that can be used to measure the different aspects that make it up, bearing in mind the large body of variables that can be applied for this purpose, as well as the need to take a long- term view. We propose a means of overcoming this problem, by applying data mining using the CHAID algorithm. This methodology enables us to create non-binary decision trees, with multiple branches for each node, providing occurrence probabilities via exclusionary rules, and is especially suitable for large sample sizes, for which, in principle, no model has yet been established for the phenomenon in question. The financial condition, in the terms defined in the present chapter, provides the characteristics necessary for such an application. The results obtained from applying the above methodology to evaluating financial independence, short-term solvency, flexibility and budgetary sustainability are highly satisfactory. The models derived, for all the Spanish local authorities analyzed, produced a success rate of over 63%, while in the case of financial independence, over 80% accuracy was achieved. Clearly, the model developed presents a high degree of explanatory and predictive capacity. For the specific cases of the worst and best values, i.e. the first and fourth quartiles for each of the elements of financial condition analyzed, an even higher rate of accuracy was recorded, ranging from 86% (for the case of the local authorities with the worst situation regarding short-term solvency) to 99.9% (for those authorities with the highest levels of financial independence). The results also suggest that the characterization of the financial condition by means of four models is a good method, as the main rules created by means of the different decision trees are made up of variables that differ depending on the element to be analyzed. In order to maintain the levels of budgetary sustainability, the most significant variables are those related to the current margin (gross savings), together with the importance of capital expenditure in the budgetary structure, while on the other hand, the short-term solvency depends on the liquid funds possessed by the entity, on the time elapsing before payments are made and received, and on the fixed charges to be met. The flexibility, however, depends on the financial load per inhabitant of the municipality, on the total sum of fixed charges, and on certain variables related to the implementation of current receipts. Finally, financial independence depends fundamentally on the transfers that the entity receives (an aspect that is predictable) and on the fiscal pressure, among other elements. It would be useful in the future to include other lines of research based on the introduction of variables concerning the social and economic context, as well as variables related to the way in which public services are managed, as these factors influence the characterization of local authorities’ financial behavior. Furthermore, it is recommended the consideration of other algorithms, within the data mining method, that could make it possible to achieve higher success rates and thus reduce the risks involved. Considering all the local authorities in question in order to classify and predict financial behavior in local government. As per from the methodological point of view, it would be more appropriate to apply other algorithms to compare the stability and prediction power of the model created, in particular, the advanced version C5.0 (Chesney, 2009), which improves how missing values are dealt with. In addition, we are aware that the automatic discretization of the continuous explanatory variables could represent a strongly impacting pre-processing statement, one that might not be necessary in certain other tree algorithms. However, since our goal is to measure the four elements of the financial condition, such an extension of the study would lead to the word limits of a book chapter being exceeded. Therefore the focusing part is on the implementation of the CHAID method to each of the above four elements, to obtain preliminary results as a starting point for future research on which we are currently working.

How to define intelligence?

Well its been a very long time since my last blog. I have been finding the answers on how do we define intelligence? And here is what I have learnt. Its been a very long time on which the debate is going on the necessities of intelligence, but sadly, there is little sign of consensus. Lets have a brief talk on it :

“AI is concerned with methods of achieving goals in situations in which the information available has a certain complex character. The methods that have to be used are related to the problem presented by the situation and are similar whether the problem solver is human, a Martian, or a computer program.”

Intelligence usually means “the ability to solve hard problems”.

“By ‘general intelligent action’ it seems to be the same sort of intelligence as we see in human action.In any real situation behavior appropriate to the ends of the system and adaptive to the demands of the environment can occur, within some limits of speed and complexity.”
And after all these years of study, we still don’t know very much about it. There are a lot more queries than answers. We all know that a well-founded definition is usually the result, rather than the starting point, of scientific research. However, there are still reasons for us to be concerned about the definition of intelligence at the current time. Though clarifying the meaning of a concept always helps communication, this problem is especially important for AI. Without a clear idea of what intelligence is, it is very hard to say why AI is different from computer science or psychology. More importantly, the researcher in this field needs to justify his/her research plan according to such a definition. Anyone who wants to work on artificial intelligence has to face a two-phase assignment:

• to choose a working definition of intelligence
• to produce it in a computer.

A working definition is a definition concrete enough that you can directly work with it. By accepting a working definition of intelligence, it does not mean that you really believe that it fully captures the concept “intelligence”, but that you will take it as a goal for your current research project. Therefore, the lack of a consensus on what intelligence is does not prevent each researcher from picking up a working definition of intelligence. Actually, the thing is, unless we keep one definition, we wont be able to claim that we are working on artificial intelligence. By accepting a working definition of intelligence, the most important commitments a researcher makes are on the acceptable assumptions and desired results, that helps us binding all the concrete work that follows. Before studying concrete working definitions of intelligence, we need to set up a general standard for what makes a definition better than others. Carnap meets the same problem when he tried to clarify the concept of “probability”. The task “consists in transforming a given more or less inexact concept into an exact one or, rather, in replacing the first by the second”, where the first may belong to everyday language or to a previous stage in the scientific language, and the second must be given by explicit rules for its use. According to him, the working definition, must fulfill the following requirements:

1. It is similar to the concept to be defined, as the latter’s vagueness permits.
2. It is defined in an exact form.
3. It is fruitful in the study.
4. It is simple, as the other requirements permit.

All these requirements are very much reasonable and suitable for the current purpose. And let us have a look what they mean concretely to the working definition of intelligence:

• Similarity: Though “intelligence” has no exact meaning in everyday language, it does have some common usages with which the working definition should agree. If we consider that a normal human beings are intelligent, but most animals and machines are either not intelligent at all or much less intelligent than human beings.
• Exactness: Given the working definition, whether a system is intelligent should be clearly decidable. For this reason, intelligence cannot be defined in terms of other ill-defined concepts, such as mind, thinking, cognition, intentionality, rationality, wisdom, consciousness, and so on, though these concepts do have close relationships with intelligence.
• Fruitfulness: The working definition should provide concrete point for the research based on it, for instance, what assumptions can be accepted, what phenomena can be ignored, what properties are desired, and so on. Most importantly, the working definition of intelligence should contribute to the solving of fundamental problems in AI.
• Simplicity: As intelligence is surely a complex mechanism, the working definition should be simple. Theoretically, a simple definition makes it possible to explore a theory in detail; and practically a simple definition is easy to implement.

For our current purpose, there is no exactly “right” or “wrong” working definition for intelligence, but there are comparative ones. When comparing proposed definitions, the four requirements may conflict. For example, one definition is more fruitful, while another is simpler, other may be exact and may be the other one is similar. In such a condition, some weighting and trade-off becomes necessary factor. However, there is no evidence showing that in general the requirements cannot be satisfied at the same time.