A recursive patterns matching model for the dynamic pattern recognition problem

Figures & data

Table 1. Matrix E_d = < S_d, C_d >.

	Ed
Sd		Cd
Signal	State	Pointer (P)^a	Weight (W)
1	False	Pointer1	[0,1]
…	…	…	[0,1]
N	False	Pointeri	[0,1]
1	False	Pointer^1
…	…	…	…
M	False	Pointer^j	[0,1]
Ud: < ΔU1, ΔU2 >

^aAn time series pointer in which we store the data.

Figure 1. Galloping horse.

Figure 2. Three levels of the hierarchy of patterns, to recognize the pattern “horse”.

Table 2. Matrix E_d = < Galloping horse>.

Ed
Sd		Cd
Signal	State	Pointer (P)^a	Weight (W)
1 =	T	Pointer^1	0.9
2 =	T	Pointer^2	0.6
3 =	T	Pointer^3	0.6
4 =	T	Pointer^4	0.6
5 =	T	Pointer^5	0.6
6 =	T	Pointer^6	0.8
7 =	T	Pointer^7	0.9
Ud: < 0.85, 0.65 >

Table 3. Physiological conditions for the emotional patterns of a car driver.

Physiological pattern of a car driver	Happy	Heart rate	From [0.1 to 0.4]
		Respiratory rate	From [0.1 to 0.4]
		Blood pressure	From [0.1 to 0.4]
	Anger	Heart rate	From [0.5 to 1.0]
		Respiratory rate	From [0.5 to 1.0]
		Blood pressure	From [0.5 to 1.0]

Table 4. Matrix Γρ_{d = angry.}

Ed
Sd		Cd
Signal	State	Pointer (P)	Weight (W)
1	F	PointerAngry1hearate	1
2	F	PointerAngry2respitarotate	0.8
3	F	PointerAngry3bloodpressure	0.9
Ud: <ΔU1, ΔU2 >

Table 5. Matrix Γρ_{d = happy.}

Ed
Sd		Cd
Signal	State	Pointer (P)	Weight (W)
1	F	PointerHappy1hearate	0.8
2	F	PointerHappy2respitatorate	0.9
3	F	PointerHappy3bloodpressure	0.8
Ud: <ΔU1, ΔU2 >

Table 6. Matrix Γρ_{d = angry.}

Ed
Sd		Cd
Signal	State	Pointer (P)	Weight (W)
1	F	PointerHappy1hearate	1
2	F	PointerHappy2respitatorate	0.9
3	F	PointerHappy3bloodpressure	0.8
Ud: <ΔU1, ΔU2 >