Download Complete Issue (pdf 3800kb) - Academic Journals

More documents

Recommendations

Info

1974 Afr. J. Pharm. Pharmacol. Figure 2. Level-spacing distribution function P(s) of SC matrix with cut-off value rcut =0.2 and rcut =0.4. Figure 3. Level-spacing distribution function P(s) of SC matrix with cut-off value rcut =0.2 and rcut =0.4. Figure 4. Number of group Ng versus cut-off value rcut by MI analysis. Black and red dots correspond to strong and weak criterion, respectively. show similar behavior as that of MI. Therefore, although the number of eigenvalues in our case is not large enough to reach the thermodynamic limit, one is still able to see a transition from a typical GOE behavior to a typical PE with increasing rcut. This means that with increasing rcut the finite network of correlations among Figure 5. Number of group Ng versus cut-off value rcut by SC analysis. Black and red dots correspond to strong and weak criterion, respectively. Figure 6. Number of group Ng versus cut-off value rcut by PC analysis. Black and red dots correspond to strong and weak criterion, respectively. the 70 symptoms also encounters a percolation transition, turning from a connected network into separate groups, which is just what we expect to see. Considering the structure of the correlation network of the symptoms, since the number of symptoms is not too large, we can directly consider the number of correlated groups the symptoms forms with increasing rcut. We used two different standards to determine whether a symptom shall be added into a group. One is a weak criterion that the symptom is added into a group if at least one symptom in that group has non-zero correlation coefficient with it. The other is a strong criterion that the symptom is added into a group only if every symptom in that group has non-zero correlation coefficient with it. Figures 4, 5 and 6 show the curve of group number Ng versus rcup.
Number of cases Pattern Shi et al. 1975 Figure 7. X-axis is pattern; Y-axis is corresponding number of cases in the data. For purple circle point, the data is symptoms data, while for green diamond point, the data is syndrome data. The percentage around each point indicates the sensitivity of the corresponding pattern. The average of them is sensitivity of the algorithm -96.48%. One can see that the curves of the two criterions by MI are quite similar, while the curves of the weak criterion by PE and SP show that the network is not split into groups at smaller value of rcut. Therefore, we may conclude that MI analysis is easier to obtain the form of uncorrelated groups. Finally we evaluate clinical meaning of the clusters by two avenues. The first avenue is using clinical theory of TCM and expert domain knowledge to estimate pattern results. As shown in Figure 7, the pattern results recognized by MI analysis (green diamond points) are highly accordant with the results diagnosed by TCM physicians (purple circle point). The second avenue is to objectively estimate each pattern discovered by above algorithm using a supervised validation method of the following three steps. Step 1. Each pattern S is returned to the unsupervised data, if all variables of the pattern simultaneously appear (their values are non-zero) on a patient, then serial number of the patient is recorded as shown in Table 1. All serial numbers are stored in a vector with Ls dimensions denoted as Vs. Step 2. We track the vectors back to the syndrome data by adding up the vectors one by one to generate a new vector. The maximal number in the new vector and the corresponding syndrome are stored as shown in Table 1. Step 3. The accuracy of a pattern is defined as the ratio of maximal number and total number, and the accuracy of the algorithm is determined by averaging the accuracy of all patterns. We found that the pattern results of MI analysis reached a high accuracy as depicted in Figure 8. The optimal rcut is 0.04. and the corresponding accuracy is 91.04%.
Page 1 and 2:
African Journal of Pharmacy and Pha
Page 3 and 4:
Editors Sharmilah Pamela Seetulsing
Page 5 and 6:
Electronic submission of manuscript
Page 7 and 8:
Fees and Charges: Authors are requi
Page 9 and 10:
Table of Contents: Volume 6 Number
Page 11 and 12:
Page 13 and 14:
Minimum inhibitory concentration (M
Page 15 and 16:
Kim et al. 1887 Table 4. Antimicrob
Page 17 and 18:
Klevens RM, Morrison M A, Nadle J,
Page 19 and 20:
Figure 1. Chemical structure of ten
Page 21 and 22:
ng/ml by 4.9 ml blank plasma with 5
Page 23 and 24:
Figure 5. Chromatogram of plasma sp
Page 25 and 26:
Figure 7. Chromatogram of a healthy
Page 27 and 28:
Figure 11. Chromatograms of standar
Page 29 and 30:
Page 31 and 32:
Shatoor 1903 Figure 1. The effect o
Page 33 and 34:
Heart rate (Beats/min.) 345 330 315
Page 35 and 36:
Table 2. Percent of changes in hear
Page 37 and 38:
heart rate on coronary flow and car
Page 39 and 40:
Urinary tract infections are the mo
Page 41 and 42:
Table 1. Study characteristics of i
Page 43 and 44:
Table 1. contd. Sun GQ 37 007 Ross
Page 45 and 46:
Study or Subgroup Cai SF 2003 Gao H
Page 47 and 48:
Study or Subgroup Cai SF 2003 Gao H
Page 49 and 50:
cure rate and clinical effective ra
Page 51 and 52: African Journal of Pharmacy and Pha
Page 53 and 54: Table 1. The expression of KAI1 and
Page 57 and 58: Figure 2. Typical chromatograms of
Page 59 and 60: Excretion studies Figure 4 shows th
Page 63 and 64: Table 1. Statistical results of the
Page 65 and 66: ethinylestradiol /100 g levonorgest
Page 67 and 68: HO CH 2OH O OH OH OCH 2CH 2 OH Figu
Page 69 and 70: A B Figure 3. Bax positive cells of
Page 73 and 74: Table 1. Summary of specific reacti
Page 75 and 76: Figure 3. Effects of the sap of Jat
Page 79 and 80: Figure 1. A schematic diagram showi
Page 81 and 82: Chen et al. 1953 Figure 4. Concentr
Page 83 and 84: Figure 6. Representative photomicro
Page 85 and 86: shock. The features of PHC can expl
Page 87 and 88: Administration of single or repeate
Page 89 and 90: Table 1. Effects of green tea on SO
Page 91 and 92: Ozardali I, Bitiren M, Ali ZK, Zeri
Page 93 and 94: EXPERIMENTAL PROCEDURE Materials an
Page 95 and 96: Scavenging rate (%) Figure 1 Scaven
Page 97 and 98: A D Figure 5 Figure Effects 5. of E
Page 101: Figure 1. Level-spacing distributio
Page 107 and 108: sodium was then dissolved in 1000 m
Page 109 and 110: Concentration (ppm) Concentration (
Page 111 and 112: various contagious diseases such as
Page 113 and 114: Table 3. Antibacterial MIC (µg/ml)
Page 115 and 116: Table 5. Contd. 1/5 14.7 ± 0.6 39.
Page 117 and 118: UPCOMING CONFERENCES 6th European C
Page 119: African Journal of Pharmacy and Pha
show all

Download Complete Issue (pdf 3800kb) - Academic Journals

Create successful ePaper yourself

Delete template?

Save as template?