Zone analysis: a visualization framework for classification problems

There have been large attempts to adopt the bias-variance framework from the regression problems to the classification problems. However, recently, it has been shown that only non-straightforward extensions exist for classification problems. In this paper, we present an alternative visualization framework for classification problems called zone analysis. Our zone analysis framework partly extends the bias-variance idea; instead of decomposing an error into two parts, i.e. the biased and unbiased components, our framework decomposes the error into K components. While bias-variance information is still contained in our framework, our framework provides interesting observations which are not obviously seen in the previous bias-variance framework, e.g. a prejudice behavior of the bagging algorithm to various unbiased instances. Our framework is suitable for visualizing an effect of context changes on learning performance. The type of context changes which we primarily investigate in the paper is “a change from a base learner to an ensemble learner such as bagging, adaboost, arc-x4 and multi-boosting”.     

Effect of fertilizer placement on nitrogen loss from sugarcane in tropical Queensland

This paper reports on the fate of nitrogen (N) in a first ratoon sugarcane (Saccharum officinarum L.) crop in the wet tropics of Queensland when urea was either surface applied or drilled into the soil 3–4 days after harvesting the plant cane. Ammonia volatilization was measured with a micrometeorological method, and fertilizer N recovery in plants and soil, to a depth of 140 cm, was determined by mass balance in macroplots with ¹⁵N labelled urea 166 and 334 days after fertilizer application. The bulk of the fertilizer and soil N uptake by the sugarcane occurred between fertilizing and the first sampling on day 166. Nitrogen use efficiency measured as the recovery of labelled N in the plant was very low. At the time of the final sampling (day 334), the efficiencies for the surface and subsurface treatments were 18.9% and 28.8%, respectively. The tops, leaves, stalks and roots in the subsurface treatment contained significantly more fertilizer N than the corresponding parts in the surface treatment. The total recoveries of fertilizer N for the plant-trash-soil system on day 334 indicate significant losses of N in both treatments (59.1% and 45.6% of the applied N in the surface and subsurface treatments, respectively). Drilling the urea into the soil instead of applying it to the trash surface reduced ammonia loss from 37.3% to 5.5% of the applied N. Subtracting the data for ammonia loss from total loss suggests that losses by leaching and denitrification combined increased from 21.8% and 40.1% of the applied N as a result of the change in method of application. While the treatment resulted in increased denitrification and/or leaching loss, total N loss was reduced from 59.1% to 45.6%, (a saving of 13.5% of the applied N), which resulted in an extra 9.9%of the applied N being assimilated by the crop. This revised version was published online in November 2006 with corrections to the Cover Date.     

Fate of urea nitrogen applied to a banana crop in the wet tropics of Queensland

This paper reports a study in the wet tropics of Queensland on the fate of urea applied to a dry or wet soil surface under banana plants. The transformations of urea were followed in cylindrical microplots (10.3 cm diameter × 23 cm long), a nitrogen (N) balance was conducted in macroplots (3.85 m × 2.0 m) with ¹⁵N labelled urea, and ammonia volatilization was determined with a mass balance micrometeorological method. Most of the urea was hydrolysed within 4 days irrespective of whether the urea was applied onto dry or wet soil. The nitrification rate was slow at the beginning when the soil was dry, but increased greatly after small amounts of rain; in the 9 days after rain 20% of the N applied was converted to nitrate. In the 40 days between urea application and harvesting, the macroplots the banana plants absorbed only 15% of the applied N; at harvest the largest amounts were found in the leaves (3.4%), pseudostem (3.3%) and fruit (2.8%). Only 1% of the applied N was present in the roots. Sixty percent of the applied N was recovered in the soil and 25% was lost from the plant-soil system by either ammonia volatilization, leaching or denitrification. Direct measurements of ammonia volatilization showed that when urea was applied to dry soil, and only small amounts of rain were received, little ammonia was lost (3.2% of applied N). In contrast, when urea was applied onto wet soil, urea hydrolysis occurred immediately, ammonia was volatilized on day zero, and 17.2% of the applied N was lost by the ninth day after that application. In the latter study, although rain fell every day, the extensive canopy of banana plants reduced the rainfall reaching the fertilized area under the bananas to less than half. Thus even though 90 mm of rain fell during the volatilization study, the fertilized area did not receive sufficient water to wash the urea into the soil and prevent ammonia loss. Losses by leaching and denitrification combined amounted to 5% of the applied N.