<p>This paper studies the identification of an <InlineEquation ID="IEq1"> <EquationSource Format="TEX">\(\mathbb {R}^d\)</EquationSource> <EquationSource Format="MATHML"><math> <msup> <mrow> <mi mathvariant="double-struck">R</mi> </mrow> <mi>d</mi> </msup> </math></EquationSource> </InlineEquation>-valued diffusion <i>X</i> when a running function of it, say <InlineEquation ID="IEq2"> <EquationSource Format="TEX">\(h(X_t)\)</EquationSource> <EquationSource Format="MATHML"><math> <mrow> <mi>h</mi> <mo stretchy="false">(</mo> <msub> <mi>X</mi> <mi>t</mi> </msub> <mo stretchy="false">)</mo> </mrow> </math></EquationSource> </InlineEquation>, is observed. A point-wise observation of the process (in other words, observing <InlineEquation ID="IEq3"> <EquationSource Format="TEX">\(h(X_t)\)</EquationSource> <EquationSource Format="MATHML"><math> <mrow> <mi>h</mi> <mo stretchy="false">(</mo> <msub> <mi>X</mi> <mi>t</mi> </msub> <mo stretchy="false">)</mo> </mrow> </math></EquationSource> </InlineEquation> in isolation) cannot identify <InlineEquation ID="IEq4"> <EquationSource Format="TEX">\(X_t\)</EquationSource> <EquationSource Format="MATHML"><math> <msub> <mi>X</mi> <mi>t</mi> </msub> </math></EquationSource> </InlineEquation> unless the <i>h</i> is injective. However observing <InlineEquation ID="IEq5"> <EquationSource Format="TEX">\(h(X_s)\)</EquationSource> <EquationSource Format="MATHML"><math> <mrow> <mi>h</mi> <mo stretchy="false">(</mo> <msub> <mi>X</mi> <mi>s</mi> </msub> <mo stretchy="false">)</mo> </mrow> </math></EquationSource> </InlineEquation> on a small interval <InlineEquation ID="IEq6"> <EquationSource Format="TEX">\([t,t+\varepsilon ]\)</EquationSource> <EquationSource Format="MATHML"><math> <mrow> <mo stretchy="false">[</mo> <mi>t</mi> <mo>,</mo> <mi>t</mi> <mo>+</mo> <mi>ε</mi> <mo stretchy="false">]</mo> </mrow> </math></EquationSource> </InlineEquation> can be enough to determine <InlineEquation ID="IEq7"> <EquationSource Format="TEX">\(X_t\)</EquationSource> <EquationSource Format="MATHML"><math> <msub> <mi>X</mi> <mi>t</mi> </msub> </math></EquationSource> </InlineEquation> exactly. The paper contain results that expand on this idea; in particular, a property of ‘fine total asymmetry’ of twice continuously differentiable <i>h</i> is introduced that depends on the fine topology of potential theory and that is both necessary and sufficient for <i>X</i> to be adapted to a natural right-continuous filtration generated by the observations. This particular filtration, though augmented with null sets, does not depend on the distribution of <InlineEquation ID="IEq8"> <EquationSource Format="TEX">\(X_0\)</EquationSource> <EquationSource Format="MATHML"><math> <msub> <mi>X</mi> <mn>0</mn> </msub> </math></EquationSource> </InlineEquation>. For real-analytic <i>h</i> the property reduces to simple asymmetry; that is, there is no nontrivial affine isometry <InlineEquation ID="IEq9"> <EquationSource Format="TEX">\(\kappa \)</EquationSource> <EquationSource Format="MATHML"><math> <mi>κ</mi> </math></EquationSource> </InlineEquation> on <InlineEquation ID="IEq10"> <EquationSource Format="TEX">\(\mathbb {R}^d\)</EquationSource> <EquationSource Format="MATHML"><math> <msup> <mrow> <mi mathvariant="double-struck">R</mi> </mrow> <mi>d</mi> </msup> </math></EquationSource> </InlineEquation> such that <InlineEquation ID="IEq11"> <EquationSource Format="TEX">\(h = h \circ \kappa \)</EquationSource> <EquationSource Format="MATHML"><math> <mrow> <mi>h</mi> <mo>=</mo> <mi>h</mi> <mo>∘</mo> <mi>κ</mi> </mrow> </math></EquationSource> </InlineEquation>. A second result concerns the case where <InlineEquation ID="IEq12"> <EquationSource Format="TEX">\(X_0\)</EquationSource> <EquationSource Format="MATHML"><math> <msub> <mi>X</mi> <mn>0</mn> </msub> </math></EquationSource> </InlineEquation> is given and <i>h</i> is merely Borel; then <i>X</i> is adapted to an augmented filtration generated by the observation process <InlineEquation ID="IEq13"> <EquationSource Format="TEX">\((h(X_t))_{t\ge 0}\)</EquationSource> <EquationSource Format="MATHML"><math> <msub> <mrow> <mo stretchy="false">(</mo> <mi>h</mi> <mrow> <mo stretchy="false">(</mo> <msub> <mi>X</mi> <mi>t</mi> </msub> <mo stretchy="false">)</mo> </mrow> <mo stretchy="false">)</mo> </mrow> <mrow> <mi>t</mi> <mo>≥</mo> <mn>0</mn> </mrow> </msub> </math></EquationSource> </InlineEquation> if <i>h</i> is ‘locally invertible’ on a subset of <InlineEquation ID="IEq14"> <EquationSource Format="TEX">\(\mathbb {R}^d\)</EquationSource> <EquationSource Format="MATHML"><math> <msup> <mrow> <mi mathvariant="double-struck">R</mi> </mrow> <mi>d</mi> </msup> </math></EquationSource> </InlineEquation> dense in the fine topology on <InlineEquation ID="IEq15"> <EquationSource Format="TEX">\(\mathbb {R}^d\)</EquationSource> <EquationSource Format="MATHML"><math> <msup> <mrow> <mi mathvariant="double-struck">R</mi> </mrow> <mi>d</mi> </msup> </math></EquationSource> </InlineEquation>.</p>

错误:搜索内容不能为空,请输入英文关键词
错误:关键词超出字数限制,请精简
高级检索

The identification of diffusions from imperfect observations

  • J. M. C. Clark,
  • Dan Crisan

摘要

This paper studies the identification of an \(\mathbb {R}^d\) R d -valued diffusion X when a running function of it, say \(h(X_t)\) h ( X t ) , is observed. A point-wise observation of the process (in other words, observing \(h(X_t)\) h ( X t ) in isolation) cannot identify \(X_t\) X t unless the h is injective. However observing \(h(X_s)\) h ( X s ) on a small interval \([t,t+\varepsilon ]\) [ t , t + ε ] can be enough to determine \(X_t\) X t exactly. The paper contain results that expand on this idea; in particular, a property of ‘fine total asymmetry’ of twice continuously differentiable h is introduced that depends on the fine topology of potential theory and that is both necessary and sufficient for X to be adapted to a natural right-continuous filtration generated by the observations. This particular filtration, though augmented with null sets, does not depend on the distribution of \(X_0\) X 0 . For real-analytic h the property reduces to simple asymmetry; that is, there is no nontrivial affine isometry \(\kappa \) κ on \(\mathbb {R}^d\) R d such that \(h = h \circ \kappa \) h = h κ . A second result concerns the case where \(X_0\) X 0 is given and h is merely Borel; then X is adapted to an augmented filtration generated by the observation process \((h(X_t))_{t\ge 0}\) ( h ( X t ) ) t 0 if h is ‘locally invertible’ on a subset of \(\mathbb {R}^d\) R d dense in the fine topology on \(\mathbb {R}^d\) R d .