• No se han encontrado resultados

3.6. Algoritmos de clasificación utilizados en EEG

3.6.1. Redes neuronales en el aprendizaje máquina

Transport behaviour is inextricably linked to the built environment. As such, urban form, transport behaviour, levels of physical activity and public health are interlinked (Saelens et al., 2003a). There is interest in the health field as to how aspects of the urban environment affect levels of physical activity, with transport patterns being an intermediary variable (Booth et al., 2001; King, 1994). Sensible urban design should include planning for people and their health (National Heart Foundation, 2004).

A number of researchers argue that efforts to make our urban environment more walkable (and cycle-friendly) are consistent with a drive to re-link public health with urban planning (Coburn, 2004; Frank et al., 2003; Greenberg et al., 1994; Greenberg et al., 2005). It is a link, they argue, that was once important, but has diminished under conventional planning paradigms. What aspects of design are therefore beneficial for people’s health, or conversely, are to the detriment of people’s health? Unfortunately, there is as yet insufficient evidence to support a causal connection between the built environment and levels of physical activity (Transportation Research Board, 2005).

Nevertheless, a study in Atlanta, US found significantly lower rates of obesity among residents of compact, dense, more pedestrian-oriented neighbourhoods with good access

implication is that residents of the former type of neighbourhood engage in more physical activity than their counterparts in the latter. Similarly, another US study found that residents of a highly walkable neighbourhood engaged in roughly 70 minutes more moderate to vigorous exercise per week, than residents of a comparatively low walkability neighbourhood (Saelens et al., 2003a). Moudon and colleagues (2006) found that higher residential densities, smaller block sizes and shorter travel distances from homes to food facilities were environmental variables positively associated with walking behaviour that was sufficient to meet health recommendations. Moreover, Perth-based research by McCormack and colleagues (2007) found that the provision of a mix of transport-related destinations, such as grocery stores sited close to residents is likely to encourage transport-related physical activity. These environmental characteristics should therefore feature in ‘walkable’ neighbourhoods.

A cross-sectional comparative study of St Louis, Missouri (relatively low walkability) and Savannah, Georgia (relatively high walkability) found associations between levels of physical activity and environmental characteristics (Hoehner et al., 2005). The researchers concluded that modifications to the environment (i.e. improving walkability) might influence the travel behaviour of the persons exposed to the changes.

Increases in activity intensity as a specific design response are argued to benefit people’s health in a number of ways. Local provision of services and facilities may promote more walking and cycling, leading to health benefits from physical activity (Frank et al., 2004). Increased permeability as part of compact neighbourhood design is conducive to physical activity and hence, can have lasting benefits for public health.

People are more likely to be moderately physically active when there are a good mix of

nearby destinations and direct routes to get there (Frank, 2000a; Frank et al., 2005;

Powell et al., 2003). Similarly, Greenberg et al. (2005: p90) argued that “making immediate neighbourhoods more walkable and bikeable is important to increase the likelihood of additional exercise”.

Conversely, when there are significant distances between destinations, walking and cycling are both time consuming and unattractive mode choices (McIndoe et al., 2005).

It follows that the existing built environment is likely to be contributing to rising levels of obesity, as sprawl makes it unfeasible for people to walk (or even cycle) to destinations (Frank et al., 2003; Frank et al., 2004; Vandegrift and Yoked, 2004).

Other research has identified the importance of attractive recreational facilities being provided locally (Active Living Research, 2004; Giles-Corti and Donovan, 2002).

Furthermore, such facilities should support multiple uses and not just cater for organised sports (Giles-Corti and Donovan, 2002). A study in Missouri, America found that people living in areas lacking outdoor recreation facilities are more likely to be overweight than counterparts in areas with a relatively high provision of facilities (Catlin et al., 2003). These findings suggest provision of quality recreation opportunities can facilitate increased physical activity, even if people drive to them.

Other research has found higher levels of moderate level utilitarian physical activity by residents of highly walkable neighbourhoods, relative to residents of lowly walkable neighbourhoods (Saelens et al., 2003a). The researchers found no significant difference, however, in vigorous and self-reported leisure activity. Similarly, Handy

between residents of highly and lowly walkable neighbourhoods. These findings illustrate how the built form may exert more of an influence on utilitarian rather than leisure activity. More research is needed, however, to support this theory.

Some other findings confuse the relationship between neighbourhood design, physical activity and health indicators. US research found no association between urban form and levels of obesity at the neighbourhood scale (Ewing et al., 2003). This would seem to reflect the variability of urban form over space and that people conduct much of their business outside of their immediate neighbourhood. Moreover, it raises questions about how successful local design and piecemeal development can be at increasing levels of physical activity.

In Perth, relatively higher density neighbourhoods outside of the CBD featuring a good mix of land uses and a well-connected street network tend to be linked to lower density, conventionally-designed neighbourhoods. Consequently, once a person leaves their compact neighbourhood they enter sprawled areas of the city and tend to have to rely on their cars for travel. It is therefore important to identify where LNs are being developed and how they are integrated with adjoining communities.

Other research has found that aspects of the physical environment are important for determining how likely people are to walk or cycle, but that such aspects are secondary to individual and social determinants of behaviour (Giles-Corti and Donovan, 2002;

Stahl et al.; 2001). Furthermore, Sallis and Owen (1999) note that health researchers have examined numerous psychological and social variables that may be associated

with a person’s travel behaviour and that this in turn can have an effect on how much physical activity they get.

The discussion has identified that the relationships between the built form, transport, physical activity and health are complex. Nevertheless, much of the research suggests car dependent, sprawled environments provide people with fewer opportunities for physical activity and consequent health benefit than compact, highly walkable environments. Significantly, however, the compact, highly walkable neighbourhoods need to be coordinated to facilitate the greatest benefits.