OPEN ACCESS
The purpose of this work is to develop an online studentteacher mentoring algorithm using statistical data based signal processing. In this approach, it is assumed that teachers’ data and attendances of the students are available in the online portal. Students will give mentoring feedback marks for the teachers according to their choices. This statistical information is transferred to discrete time signal by scaling and shifting operation. Performance of the algorithm is established by determining average mentoring feedback value, standard deviation, signaltonoise ratio (SNR), coefficient variance and typical error. Statistical signal processing algorithm is used to compute the whole mentoring system and error in algorithm is calculated based on SNR. Here, 30 students are grouped under 5 teachers equally for mentoring and using the proposed algorithm achieved typical error varies from about 11% to 12.5%. The findings of this investigation will be useful for online studentmentoring system in educational organizations for the enhancement of quality of education.
online feedback, studentteacher mentoring, mentoring algorithm, statistical signal processing
In order to increase daytoday interaction with the teachers, student mentoring is an important component in an educational institute. A mentoring is a process in organizations which reduces the gap between mentees and mentor and built the bright career, better life style and will make society prosperous and peace. In this paper, a new statistical data based signal processing algorithm is used for online studentteacher mentoring system in educational organizations. In this approach, it is assumed that teachers’ data and attendances of the students are available in the online portal. Total number of students is grouped under different tutormentors. Teachers will mentor the students for different activities, like, education, attendances in class, higher studies and other extracurricular activities. Streamwise all subjects name with code number, teachers’ name subjectwise, semesterwise groups of students name and roll numbers with particular faculties are available in the data base system of the organization. Students will give online feedback for his/her mentor according to his/her choice, but student attendance will be considered to compute mentoring feedback as one of the parameters for marking system. This statistical information is transferred to discrete time signal by scaling and shifting operation. The algorithm is developed in such a manner that those who have less than (7585) % attendances only 50% mentoring points will generate to give the mentoring feedback to their concern teacher, (8695) % attendances only 75% mentoring points will generate to give the mentoring feedback to their concern teacher and equal and above 95% attendances 100% mentoring points will generate to give the mentoring feedback to their concern teacher respectively. Performance of the algorithm is established by determining average mentoring feedback value, standard deviation, signaltonoise ratio (SNR), coefficient variance and typical error. Statistical signal processing algorithm is used to compute the whole mentoring system and error in algorithm is calculated based on SNR. Here, 30 students are grouped under 5 teachers (that is, 6 students under each teacher) for mentoring and using the proposed algorithm achieved typical error varies from about 11% to 12.5%. The findings of this investigation will be useful for online studentmentoring system in educational organizations for the enhancement of quality of education.
Mentoring is nowadays very important in education area of organization to interactions between the teachers (mentors) and the students (mentees) built students’ future carrier, life and minimize uncivilian works in society. A mentoring is a process in organizations which reduces the gap between mentees and mentor and built the bright career, better life style and will make society prosperous and peace. Mentoring programs can be valuable for mentor in education organization and commercial purpose to encourage the mentees in selfconfidence to lifelong positive professional impacts, such as learning and understanding by accepting feedback from mentors. So mentors have an opportunity to share their knowledge, strengthen their interpersonal technique skills to make the mentees a good social person in society.
Overall learning and attitude of a student are very much dependent on effective mentoring of teachers (Anderson, 1991; Wortmann et al., 2008; Deans et al., 2009; Wanberg et al., 2006; Block and Georgiadis, 2012; Xiao, 2013). Because of advancement of information and communication technology, online mentoring is gaining much more attention worldwide (Sohn et al., 2012; Srivichai et al., 2012; Quintana and Zambrano, 2014; Rhodes and Spencer, 2010), though facetoface communication cannot be denied for mentormentee communication. The models for mentormentee communications, reported earlier might not be useful for online mentoring. Therefore, some common basics of mentormentee communication for online mentoring is discussed in (Nuankaew and Temdee, 2015). Positive impacts on teaching quality, teachers’ morality and teacherstudent interactions are observed by using a computer joystick (Pennings et al., 2014). The principles (Hendrickxa et al., 2015) underpinning efficacious inclusive teacher education, research project and recommended for the context of an organizations including experience of teachers. The teacher (Sedova et al. 2016) should be dialogic teaching, classroom discourse parameters was identified, the amount of talk with reasoning increased and other indicators of dialogic teaching. From qualitative statistical data on teachers’ career it is found that teacher attrition should be more complex and nonlinear phenomenon (Lindqvist et al., 2014). The quality of studentteacher interaction is related to teacher and student based on instructional support, positive emotional relationship with students during the education (Krstić, 2015).
There are several methods for the assessment of different aspects of teacherstudent socialemotional relationship, interaction inside and outside the class room but very few literatures are available on online studentmentoring system. In this paper, based on the literature review, a perfect and automatic feedback mentoring system is developed, which is applicable for average weighted based attendance of mentees. This will give the desired feedback marks to their concern mentors according their average weighted based attendance. Feedback mentoring improvement based on average weighted attendance of mentees is adapted from several related scales and factor from several research papers. Also according to their mentoring feedback, the mentoring marks will be entered against the respective mentors and the proposed method takes care that only after completion of mentees feedback admit card will generate to appear in the final examination. For the implementation of this mentoring feedback system, a signal processing algorithm is used based on statistical data for the mentors. Using statistical data, mean, standard deviation, signaltonoise ratio (SNR), typical error and coefficient variance are computed by using signal processing. MATLAB is used for simulation.
The rest of the paper is organized as: Section 2 is the proposed online mentoring algorithm, Section 3 describes the implementation of the proposed statistical data based signal processing algorithm for online mentoring including performance of proposed algorithm, Section 4 is conclusion and future work and last section is references.
The algorithm, proposed for online mentoring feedback system, is described in detail below.
Input in Institute Portal: Insert semesters; Insert branch names; Teachers’ names group wise roll nos.;
Input in Institute Portal: Subjects name with subject codes; Admit card format;
Input in Institute Portal: Roll numbers of students with percentage of attendance;
Input in Institute Portal: Feedback points [For each mentoring point answer bits have, (a) = 0.0 mark, (b) = 0.25 mark, (c) = 0.75 mark and (d) =1.0 mark];
Input by Students: Student’s ID number and Password
Estimation of Mentoring: Statistical databased signal processing algorithm will estimate mentoring marks and teacher’sname wise
Output: Sum of all mentoring marks teacher’sname wise.
Output: Admit card will be generating for eligible students (attendance ≥75%) after inserting their mentoring feedback on the portal. No admit card will be generated for the eligible students who have not completed online feedback process.
The stepwise procedure of the proposed algorithm for online mentoring is described below:
Step1: Streamwise all subjects name with code number, teachers’ name subjectwise, semesterwise groups of students name and roll numbers with particular faculties are available in the data base system of the organization. Mentees roll nos. (1 6) for Prof. A, roll nos. (7 12) for Prof. B, roll nos. (1318) for Prof. C, roll nos. (1924) for Prof. D and roll nos. (2530) for Prof. E. 20 numbers of mentoring points with multiple choices as (a) = 0.0 mark, (b) = 0.25 mark, (c) = 0.75mark and (d) = 1.0 mark are uploaded in organization database.
Step2: Teachers will update attendance in the organization data base system immediately after taking the classes. After completion of course total attendance of the students will be stored in the data based system. The students roll nos. are divided in three tier such as less than (7585) % attendances, (8695) % attendances and equal and above 95% attendances to give the feedback about the mentoring to their concern teacher respectively. Those who have less than 75% attendances, their roll numbers will not be available in the data base of organization.
Step3: After giving online mentoring feedback for their concern mentoring teacher by the students and feedback marks about the mentoring will be entered automatically against the respective mentor’s name wise.
Step4: The algorithm is developed in such a manner that those who have less than (7585) % attendances only 50% mentoring points will generate to give the mentoring feedback to their concern teacher, (8695) % attendances only 75% mentoring points will generate to give the mentoring feedback to their concern teacher and equal and above 95% attendances 100% mentoring points will generate to give the mentoring feedback to their concern teacher respectively.
Step5: After completion of the process of mentoring feedback by the students, desired mentoring feedback marks will be enrolled in respective teachers format and admit card will be generated for the student to appear in the examination. In this process the concern teachers will be evaluated correctly by the eligible students.
Step6: Then such statistical information will be transferred to discrete time signal by scaling and shifting operation and will be simulated by statistical signal processing to get accurate information. From the statistical information the average mentoring feedback value, standard deviation, signal to noise ratio (SNR), coefficient variance and typical error will be determined.
Good statistical signal representations (e.g. timefrequency or timescale) are required for many engineering applications (Smith, 1997). Signal is described how can one parameter is related to another parameter. In feedback system, feed marks of different teachers are collected which are represented in Yaxis and roll numbers of different students are represented in Xaxis. These parameters are also designated in Yaxis and Xaxis as amplitude and time of signal respectively. So, the horizontal axis represented the other parameter of the signal which is the independent variable. Time is a common parameter which is represented in horizontal axis. In generally, the horizontals axis represented as sampler number. The parameters in Yaxis (dependent variable) are represented as function of the parameter on the Xaxis (the independent variable). This independent variable represents, how and where each samples is considered, while the dependent variable is actual measurement, like a twoterminal source emitting mutually correlated signals in X & Y axis respectively (Amari and Han, 1989). Mathematical models serve as tools in the analysis and designing signal from statistical data analysis of mentoring feedback marks into delta function. Discrete continuous signals can be decomposed into scaled and shifted delta functions comparing with mentoring feed marks of teachers given by students as well as students roll numbers. Statistical data (Rosen and Porat, 1989) were sensitive to deviations of the data from the ideal model as measurement of noise, temporary deviation from stationary (change in amplitude etc) & outliers. Depending on the specify application, the algorithm may also have to be adaptive to changing signal and noise environments.
The mean is designated as “µ” in statistics for the average value of a signal. If there are X(n) numbers of signal with N numbers of discrete times, then mean is calculated as (Rosen and Porat, 1989)
$\mu=\frac{1}{N} \sum_{n=0}^{N1} X(n)$ (1)
where X (n) represents the mentoring value of the signal at particular time & N represent the total sum of times where n=0,1,2......N. The calculation is completed by dividing the sum by N.
Signals may be represented as discrete versions or repetitive wave forms as sine or square with peak to peak amplitude. But such type of signals is referred as in random nature. Deviation of N sample from mean value is described by standard deviation (σ). The average deviation of a signal is found by summing the deviations of all the individual samples and then divided by the number of samples N. The average deviation gives a single number representing the typical distance which samples are mean, while the average deviation is almost never used in statistics. So it is not well fit with physics that how can the signals operated. When random noise signal (Hero, 2008) combine in an electric circuit & the resultant noise is equal to the combined power of individual signals. The standard deviation is the average deviation with its power, but not in amplitude, which is done by taking the squaring each of the deviation before taking the average. Then standard deviation (σ) is represented as
$\left.\sigma=\sqrt{[ } \frac{1}{N} \sum_{n=0}^{N1}\{\mathrm{X}(\mathrm{n})\mu\}\right]$ (2)
A signal without any DC component, its r.m.s. value is same as its standard deviation (σ). Here the mean is measured while the standard deviation represented noise and other interferences. Stochastic process (Claus, 1993) of homogeneous trees aimed to process the multi scale statistical signal & gave signal level of resolution in terms of signal to noise ratio (SNR) which is (Johnson, 2013)
$S N R(\rho)=(\mu / \sigma)$ (3)
Coefficient of variance CV $(\vartheta)$ is defined as (Johnson, 2013)
$C V(\vartheta)=100(\sigma / \mu)$ (4)
The random signal and typical error (Ɛ) [TE] between mean and the N points, and mean of the process is given by
$\varepsilon=\sigma / \sqrt{N}$ (5)
Table 1. Conversion of mentoring feedback marks to discrete time signal
Mentors 
Mentorwise Representation of Discrete Time Signal 
Prof. A 
X_{1}(n)=3.75δ(n)+3.0δ(n1)+2.5δ(n2)+2.75δ(n3)+2.5δ(n4)+2.5δ(n5) 
Prof. B 
X_{2}(n)=2.5δ(n6)+3.0δ(n7)+4.0δ(n8)+2.5δ(n9)+2.0δ(n10)+3.0δ(n11) 
Prof. C 
X_{3}(n)= 2.25δ(n12)+2.5δ(n13)+3.5δ(n14)+ 3.0 δ(n15) + 2.5 δ(n16)+3.0δ(n17) 
Prof. D 
X_{4}(n)= 1.5δ(n18)+3.0 δ(n19)+3.25δ(n20)+3.0 δ(n21)+3.0δ(n22)+1.5δ(n23) 
Prof. E 
X_{5}(n)= 2.75 δ(n24)+2.75δ(n25)+ 2.25δ(n26)+ 2.5δ(n27)+ 3.0δ(n28)+1.5 δ(n29) 
Statistical data based signal processing is performed by representing discrete time signal for each feedback mentoring values as w_{1}δ(n)+w_{2}δ(n1)+w_{3}δ(n2)+ …………+w_{N }δ[n(N1)]. Where, N is the number of students, taken into consideration, δ(n) is the impulse delta function whose value is 1 for all integer values of ‘n’ and w_{N} is the weight of the signal as described in the Table 1.
The mentoring feedback marks in analog and digital forms for all the 5 professors are shown in Fig.1Fig.5.
For online feedback of 30 students different statistical parameters, estimated from statistical signal processing for different subjects are tabulated in Table 2.
Typical error in the Table 2 varies from about 11% to about 12.5%. This typical error is high in this case because error is calculated based on mean of the signal in this paper. Beyond the mean the signal noise is present over the span of standard deviation. Subjectwise teacherwise SNRs calculated from the algorithm are much higher than the coefficient variance and also from the typical error. These estimation parameters are tabulated in Table 3. In this table SNR is of the order of 5 in all the cases. If number of data is large SNR is also large.
In Table 3, SNR/CV is less than SNR/TE, that is, it shows that noise of CV is more than noise of TE. Therefore, the online feedback estimation using the proposed statistical data based signal processing is very effective for evaluation of mentoring points of teachers in an institute.
Figure 1. Mentoring feedback marks for Prof. A
Figure 2. Mentoring feedback marks for Prof. B
Figure 3. Mentoring feedback marks for Prof. C
Figure 4. Mentoring feedback marks for Prof. D
Figure 5. Mentoring feedback marks for Prof. E
Table 2. Estimated parameters for 30 students using statistical signal processing
Roll. No. 
Average Attendance In all subjects 
Ment. Feedback marks 
Mean (µ) 
Std Deviation (σ) 
SNR (ρ) 
SNR/CV 
SNR/TE 
Mentors 
01 
97.2 
3.75 
2.8 
0.62 
4.5 
0.2 
17.74 
Prof. A 
02 
87.8 
3.0 

03 
87.8 
2.5 

04 
86.8 
2.75 

05 
94.2 
2.5 

06 
87.8 
2.5 

07 
87.4 
2.5 
2.8 
0.60 
4.65 
0.22 
19.0

Prof. B 
08 
88.6 
3.0 

09 
97.4 
4.0 

10 
87.4 
2.5 

11 
80.0 
2.0 

12 
91.4 
3.0 

13 
93.0 
2.25 
2.75 
0.39 
7.1 
0.50 
44.91 
Prof. C 
14 
94.4 
2.5 

15 
97.6 
3.5 

16 
93.4 
3.0 

17 
91.6 
2.5 

18 
85.0 
3.0 

19 
80.4 
1.5 
2.83 
0.58 
4.85 
0.24 
20.32 
Prof. D 
20 
89.0 
3.0 

21 
96.8 
3.25 

22 
92.2 
3.0 

23 
95.6 
3.0 

24 
93.0 
1.5 

25 
91.6 
2.75 
2.5 
0.42 
5.98 
0.36 
35.01 
Prof. E 
26 
93.0 
2.75 

27 
90.8 
2.25 

28 
87.2 
2.5 

29 
91.0 
3.0 

30 
94.6 
1.75 
Table 3. Estimation parameters for online feedback
Estimation Parameter 
Prof. A 
Prof. B 
Prof. C 
Prof. D 
Prof. E 
SNR / CV 
25.577 
29.300 
26.656 
22.647 
23.431 
SNR / TE 
44.987 
47.059 
44.396 
38.280 
39.103 
Online mentoring feedback system proposed here is more transparent for evaluation of mentoring of a particular faculty and it will help the learning and professional development of students in the institute. In this process the eligible students will get admit card to appear the examination and also mentors will get proper justice from the students, because the feedback points are generated according the average attendance based by the students. The information will be useful to the teacher(s) as well as to the organization in the ongoing efforts to enhance the quality of education. In statistical signal processing, processing of signal is treated as stochastic processes, dealing with their statistical properties (e.g., mean, standard deviation, SNR and coefficient of variance, etc.). From the research work it is found that the better data is obtained for higher SNR and for lower value for the coefficient variance. Signal processing using Bispectrum to determine the better data for online mentoring may be the future work on this topic.
Amari S. I., Han T. S. (1989). Statistical Inference under multi terminal rate restrictions: A differential geometric approach. IEEE Transaction on Information Theory, Vol. 35, No. 2, pp. 217227. https://doi.org/10.1109/18.32118
Anderson L. W. (1991). Increasing teacher effectiveness. Fundamentals of Educational Planning, 39. United Nations Educational, Scientific, and Cultural Organization, Paris (France), Paris. https://doi.org/10.1177/019263657205636703
Block B. M., Georgiadis A. (2012). Projectmentoring in engineering education  a competenceoriented teaching and learning approach. Proceedings of the 2012 IEEE Global Engineering Education Conference (EDUCON), pp. 1–8. https://doi.org/10.1109/EDUCON.2012.6201044
Claus B. H. (1993). Multi scale statistical signal processing: Identification of a multiscale AR process from a sample of an ordinary signal. IEEE Transactions on Signal Processing, Vol. 41, No. 12, pp. 32663274. https://doi.org/10.1109/78.258072
Deans F., Oakley P., James R. (2009). Coaching and mentoring for leadership development in civil society. International NGO Training and Research Centre (INTRAC).
Hendrickxa M. M. H. G., Mainharda T., BoorKlipb, H. J., Cillessen A. H. M., Brekelmans M. (2015). Social dynamics in the class room: Teacher support and conflict and the peer ecology. Teaching and Teacher Education, Vol. 53, pp. 3040. https://doi.org/10.1016/j.tate.2015.10.004
Hero A. O. (2008). Statistical methods for signal processing. Lecture Notes.
Johnson D. H. (2013). Statistical signal processing. Lecture Notes.
Krstić K. (2015). Attachment in the studentteacher relationship as a factor of school achievement. Teaching Innovations, Vol. 28, No. 3, pp. 167–188.
Lindqvist P., Nordänger U. K., Carlsson R. (2014). Teacher attrition the first five years a multifaceted image. Teaching and Teacher Education, Vol. 40, pp. 94103. https://doi.org/10.1016/j.tate.2014.02.005
Nuankaew P., Temdee P. (2015). Of online community: Identifying mentor and mentee with compatible different attributes and decision tree. 12th Intl. Conference on Electrical Engineering/Electronics, Computer, Telecommunications and Information Technology (ECTICON). https://doi.org/10.1109/ECTICon.2015.7207130
Pennings H. J. M., Van Tartwijk J., Wubbels T. (2014). Realtime teacherstudent interactions: A dynamic systems approach. Teaching and Teacher Education, Vol. 37, pp. 183193. https://doi.org/10.1016/j.tate.2013.07.016
Quintana M. G. B., Zambrano E. P. (2014). Ementoring: The effects on pedagogical training of rural teachers with complex geographical accesses. Computers in Human Behavior, Vol. 30, pp. 629–636. https://doi.org/10.1016/j.chb.2013.07.042
Rhodes J. E., Spencer R. (2010). Structuring mentoring relationships for competence, character, and purpose. New Directions for Youth Development, Vol. 126, pp. 149–152. https://doi.org/10.1002/yd.356
Rosen Y., Porat B. (1989). Optimal ARMA parameter estimate based on the sample covariances for data with missing observations. IEEE Transaction on Information Theory, Vol. 35, No. 2, pp. 342349. https://doi.org/10.1109/18.32128
Sedova K., Sedlacek M., Svaricek R. (2016). Teacher professional development as a means of transforming student classroom talk. Teaching and Teacher Education, Vol. 57, pp. 1425. https://doi.org/10.1016/j.tate.2016.03.005
Smith S. W. (1997). The scientist and engineer's guide to digital signal processing. California Technical Publishing, pp. 503534.
Sohn M., Kwon Y. M., Lee H. J. (2012). Contextbased hybrid semantic matching framework for ementoring system. 15th International Conference on NetworkBased Information Systems (NBiS), pp. 691–696. https://doi.org/10.1109/NBiS.2012.149
Srivichai P., Meksamoot K., Jengjalern A. (2012). Integrated knowledge management system based mentoring for new university staff development. 9th International Conference on ICT and Knowledge Engineering (ICT Knowledge Engineering), pp. 150–155. https://doi.org/10.1109/ICTKE.2012.6152397
Wanberg C. R., KammeyerMueller J., Marchese M. (2006). Mentor and protégé predictors and outcomes of mentoring in a formal mentoring program. Journal of Vocational Behavior, Vol. 69, No. 3, pp. 410–423. https://doi.org/10.1016/j.jvb.2006.05.010
Wortmann K., Cavanaugh C., Kennedy K. (2008). Online teacher support programs: Mentoring and coaching models. North American Council for Online Learning.
Xiao X. (2013). The study of undergraduate mentorship system based on the fuzzy comprehensive assessment method. 2013 Fourth International Conference on Digital Manufacturing and Automation (ICDMA), pp. 1573–1576. https://doi.org/10.1109/ICDMA.2013.377