Claims:1. A processor implemented method (200) comprising:
obtaining boundary values (bi) for a plurality of signal features under consideration, the plurality of signal features being measured by a plurality of sensors, wherein the boundary values define limits for the plurality of signal features to classify associated signal features as belonging to a faulty group or a healthy group (202);
obtaining readings (si) from each of the plurality of sensors, wherein the obtained readings together represent a signal, and wherein a deviation from the boundary values towards the faulty group renders the associated readings as anomalous readings (204);
identifying the signal as a faulty deceptive signal when the signal is characterized by (i) a relatively small number of anomalous readings being obliterated by remaining non-anomalous readings from the obtained readings, wherein the relatively small number is decided empirically; and (ii) each of the remaining non-anomalous readings being ranked above a third quartile of a reading range associated with each of the plurality of signal features (206);
scaling and weighing the readings (si) to obtain a comparable distance between the plurality of signal features comprising the signal (208); and
computing a directional measure (d) representative of a distance of the signal from a classifying boundary for determining a level of severity of the faulty deceptive signal, wherein the classifying boundary is based on the obtained boundary values (210).

2. The processor implemented method of claim 1 further comprising: converting the computed directional measure (d) to a probability of severity based on an associated business domain corresponding to the signal features under consideration using regression analysis (212).

3. The processor implemented method of claim 1, wherein each of the plurality of signal features is associated with one or more boundary values.

4. The processor implemented method of claim 1, wherein the step of obtaining boundary values comprises performing statistical analysis.

5. The processor implemented method of claim 1, wherein the relatively small number is less than 20%.

6. The processor implemented method of claim 1, wherein the step of scaling and weighing the readings is based on interquartile ranges (IQR) associated with the plurality of signal features.

7. The processor implemented method of claim 1, wherein a non-negative value for the directional measure (d) renders the signal being classified as belonging to the faulty group and a negative value for the directional measure (d) renders the signal being classified as belonging to the healthy group.

8. The processor implemented method of claim 1, wherein the step of computing the directional measure (d) is based on a sum of exponentiated differences between the signal features and corresponding boundary values.

9. The processor implemented method of claim 2, wherein the step of converting the computed directional measure (d) to a probability of severity comprises mapping the computed directional measure to a pre-defined probability of severity obtained through regression analysis.

10. A system (100) comprising:
one or more data storage devices (102) operatively coupled to one or more hardware processors (104) and configured to store instructions configured for execution by the one or more hardware processors to:
obtain boundary values (bi) for a plurality of signal features under consideration, the plurality of signal features being measured by a plurality of sensors, wherein the boundary values define limits for the plurality of signal features to classify associated signal features as belonging to a faulty group or a healthy group;
obtain readings (si) from each of the plurality of sensors, wherein the obtained readings together represent a signal, and wherein a deviation from the boundary values towards the faulty group renders the associated readings as anomalous readings;
identify the signal as a faulty deceptive signal when the signal is characterized by (i) a relatively small number of anomalous readings being obliterated by remaining non-anomalous readings from the obtained readings, wherein the relatively small number is decided empirically; and (ii) each of the remaining non-anomalous readings being ranked above a third quartile of a reading range associated with each of the plurality of signal features;
scale and weigh the readings (si) to obtain a comparable distance between the plurality of signal features comprising the signal; and
compute a directional measure (d) representative of a distance of the signal from a classifying boundary for determining a level of severity of the faulty deceptive signal, wherein the classifying boundary is based on the obtained boundary values.

11. The system of claim 10, wherein the one or more hardware processors are further configured to convert the computed directional measure (d) to a probability of severity based on an associated business domain corresponding to the signal features under consideration using regression analysis.

12. The system of claim 10, wherein each of the plurality of signal features is associated with one or more boundary values.

13. The system of claim 10, wherein the one or more hardware processors are further configured to obtain boundary values by performing statistical analysis.

14. The system of claim 10, wherein the relatively small number is less than 20%.

15. The system of claim 10, wherein the one or more hardware processors are further configured to scale and weigh the readings based on interquartile ranges (IQR) associated with the plurality of signal features.

16. The system of claim 10, wherein a non-negative value for the directional measure (d) renders the signal being classified as belong to the faulty group and a negative value for the directional measure (d) renders the signal being classified as belong to the healthy group.

17. The system of claim 10, wherein the one or more hardware processors are further configured to compute the directional measure (d) based on a sum of exponentiated differences between the signal features and corresponding boundary values.

18. The system of claim 11, wherein the one or more hardware processors are further configured to convert the computed directional measure (d) to a probability of severity by mapping the computed directional measure to a pre-defined probability of severity obtained through regression analysis.
, Description:FORM 2

THE PATENTS ACT, 1970
(39 of 1970)
&
THE PATENT RULES, 2003

COMPLETE SPECIFICATION
(See Section 10 and Rule 13)

Title of invention:
DETECTION OF FAULTS IN DECEPTIVE SIGNALS AND COMPUTING SEVERITY THEREOF

Applicant:
Tata Consultancy Services Limited
A company Incorporated in India under the Companies Act, 1956
Having address:
Nirmal Building, 9th Floor,
Nariman Point, Mumbai 400021,
Maharashtra, India

The following specification particularly describes the invention and the manner in which it is to be performed.
TECHNICAL FIELD
The disclosure herein generally relates to Internet of Things (IOT), and more particularly relates to IOT-based systems and methods for detection of faults in deceptive signals and computing severity thereof.

BACKGROUND
Mission critical activities require continual or periodic monitoring of several features to detect indications of a fault occurrence. A vigilante system needs to frequently evaluate if performances of at least key features being monitored remain within acceptable limits. Typically in an Internet of Things (IOT) environment, there may be multiple sensors in operation, as many as 1000 or more, which are designed to measure responses from the key features that are considered critical for a business. General examples of continual or periodic monitoring from multiple sensors include flight control operation, measurements involving physical or information security, physiological signals of critical patients, and the like. A common aspect of these activities is to monitor situations that trigger sensor measurements to move outside of acceptable limits leading to generation of an alert instantaneously. However, when there are multiple readings from multiple sensors contributing to a signal characterized to detect health or a monitored aspect of a system being monitored, automated systems that employ machine learning need to be capable of analyzing failure situations including even a single point failure situation associated with a feature when every other feature being monitored may seem to be working as intended. Such signals may be deemed deceptive and pose a challenge for analyses and gauging severity associated thereof.

SUMMARY
Embodiments of the present disclosure present technological improvements as solutions to one or more of the above-mentioned technical problems recognized by the inventors in conventional systems.
In an aspect, there is provided a processor implemented method comprising: obtaining boundary values (bi) for a plurality of signal features under consideration, the plurality of signal features being measured by a plurality of sensors, wherein the boundary values define limits for the plurality of signal features to classify associated signal features as belonging to a faulty group or a healthy group; obtaining readings (si) from each of the plurality of sensors, wherein the obtained readings together represent a signal, and wherein a deviation from the boundary values towards the faulty group renders the associated readings as anomalous readings; identifying the signal as a faulty deceptive signal when the signal is characterized by (i) a relatively small number of anomalous readings being obliterated by remaining non-anomalous readings from the obtained readings, wherein the relatively small number is decided empirically; and (ii) each of the remaining non-anomalous readings being ranked above a third quartile of a reading range associated with each of the plurality of signal features; scaling and weighing the readings (si) to obtain a comparable distance between the plurality of signal features comprising the signal; and computing a directional measure (d) representative of a distance of the signal from a classifying boundary for determining a level of severity of the faulty deceptive signal, wherein the classifying boundary is based on the obtained boundary values.
In another aspect, there is provided a system comprising: one or more data storage devices operatively coupled to the one or more processors and configured to store instructions configured for execution by the one or more processors to: obtain boundary values (bi) for a plurality of signal features under consideration, the plurality of signal features being measured by a plurality of sensors, wherein the boundary values define limits for the plurality of signal features to classify associated signal features as belonging to a faulty group or a healthy group; obtain readings (si) from each of the plurality of sensors, wherein the obtained readings together represent a signal, and wherein a deviation from the boundary values towards the faulty group renders the associated readings as anomalous readings; identify the signal as a faulty deceptive signal when the signal is characterized by (i) a relatively small number of anomalous readings being obliterated by remaining non-anomalous readings from the obtained readings, wherein the relatively small number is decided empirically; and (ii) each of the remaining non-anomalous readings being ranked above a third quartile of a reading range associated with each of the plurality of signal features; scale and weigh the readings (si) to obtain a comparable distance between the plurality of signal features comprising the signal; and compute a directional measure (d) representative of a distance of the signal from a classifying boundary for determining a level of severity of the faulty deceptive signal, wherein the classifying boundary is based on the obtained boundary values.
In yet another aspect, there is provided a computer program product comprising a non-transitory computer readable medium having a computer readable program embodied therein, wherein the computer readable program, when executed on a computing device, causes the computing device to: obtain boundary values (bi) for a plurality of signal features under consideration, the plurality of signal features being measured by a plurality of sensors, wherein the boundary values define limits for the plurality of signal features to classify associated signal features as belonging to a faulty group or a healthy group; obtain readings (si) from each of the plurality of sensors, wherein the obtained readings together represent a signal, and wherein a deviation from the boundary values towards the faulty group renders the associated readings as anomalous readings; identify the signal as a faulty deceptive signal when the signal is characterized by (i) a relatively small number of anomalous readings being obliterated by remaining non-anomalous readings from the obtained readings, wherein the relatively small number is decided empirically; and (ii) each of the remaining non-anomalous readings being ranked above a third quartile of a reading range associated with each of the plurality of signal features; scale and weigh the readings (si) to obtain a comparable distance between the plurality of signal features comprising the signal; and compute a directional measure (d) representative of a distance of the signal from a classifying boundary for determining a level of severity of the faulty deceptive signal, wherein the classifying boundary is based on the obtained boundary values.
In an embodiment of the present disclosure, the one or more hardware processors are further configured to convert the computed directional measure (d) to a probability of severity based on an associated business domain corresponding to the signal features under consideration using regression analysis.
In an embodiment of the present disclosure, each of the plurality of signal features is associated with one or more boundary values.
In an embodiment of the present disclosure, the one or more hardware processors are further configured to obtain boundary values by performing statistical analysis.
In an embodiment of the present disclosure, the relatively small number is less than 20%.
In an embodiment of the present disclosure, the one or more hardware processors are further configured to scale and weigh the readings based on interquartile ranges (IQR) associated with the plurality of signal features.
In an embodiment of the present disclosure, a non-negative value for the directional measure (d) renders the signal being classified as belonging to the faulty group and a negative value for the directional measure (d) renders the signal being classified as belonging to the healthy group.
In an embodiment of the present disclosure, the one or more hardware processors are further configured to compute the directional measure (d) based on a sum of exponentiated differences between the signal features and corresponding boundary values.
In an embodiment of the present disclosure, the one or more hardware processors are further configured to convert the computed directional measure (d) to a probability of severity by mapping the computed directional measure to a pre-defined probability of severity obtained through regression analysis.
It is to be understood that both the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the embodiments of the present disclosure, as claimed.

BRIEF DESCRIPTION OF THE DRAWINGS
The accompanying drawings, which are incorporated in and constitute a part of this disclosure, illustrate exemplary embodiments and, together with the description, serve to explain the disclosed principles.
FIG.1 illustrates an exemplary block diagram of a system for detection of faults in deceptive signals and computing severity thereof, in accordance with an embodiment of the present disclosure.
FIG.2 is an exemplary flow diagram illustrating a computer implemented method for detection of faults in deceptive signals and computing severity thereof, in accordance with an embodiment of the present disclosure.
FIG.3 illustrates variation of directional measure, in accordance with an embodiment of the present disclosure, depicted through a representative line for readings from a single sensor and its boundary value.
FIG.4 illustrates fitment of the computed directional measure (d) to the probability of severity based on an associated business domain, in accordance with an embodiment of the present disclosure.
It should be appreciated by those skilled in the art that any block diagram herein represent conceptual views of illustrative systems embodying the principles of the present subject matter. Similarly, it will be appreciated that any flow charts, flow diagrams, state transition diagrams, pseudo code, and the like represent various processes which may be substantially represented in computer readable medium and so executed by a computing device or processor, whether or not such computing device or processor is explicitly shown.
DETAILED DESCRIPTION
Exemplary embodiments are described with reference to the accompanying drawings. In the figures, the left-most digit(s) of a reference number identifies the figure in which the reference number first appears. Wherever convenient, the same reference numbers are used throughout the drawings to refer to the same or like parts. While examples and features of disclosed principles are described herein, modifications, adaptations, and other implementations are possible without departing from the spirit and scope of the disclosed embodiments. It is intended that the following detailed description be considered as exemplary only, with the true scope and spirit being indicated by the following claims.
In the context of the present disclosure, the expressions “problem” or “fault” or “failure” conditions relate to scenarios or conditions that require attention or treatment or a response from associated responsible authorities. For instance, a problem condition can be a medical condition of a patient, faulty or emergency conditions of a flight, unacceptable conditions resulting from assessments in offering jobs, admissions or degrees, and the like. In the context of the present disclosure, a “problem” or “fault” or “failure” condition may not necessarily be associated with a fault but it may refer to one or mutually exclusive binary conditions that is of interest for a purpose. Readings from multiple sensors monitoring various features of a monitored system may be provided as a signal for analyzing the health of the monitored system. Systems and methods of the present disclosure address specifically signals referred hereinafter as deceptive signals that conventional machine learning algorithms find difficult to analyze or classify as a fault condition. Difficulty arises particularly in a condition where most of the features constituting the signal are associated with readings that are within acceptable limits along with a very small number of features (one or two) that may be associated with readings outside acceptable limits leading to the overall health of the signal being mistakenly inferred as healthy and acceptable. The expression “deceptive signals” in the context of the given disclosure refers to such problematic signals that are characterized by (i) a relatively small number of anomalous readings being obliterated by remaining non-anomalous readings from the readings obtained from associated sensors; and (ii) each of the remaining non-anomalous readings being ranked above a third quartile of a reading range associated with each of the plurality of signal features. In an embodiment, the relatively small number may be decided empirically. The expression “anomalous readings” in the context of the present disclosure refers to readings that do not comply with acceptable limits associated with the feature being monitored. Accordingly, the deceptive signals may appear healthy and very far away from a group separation boundary as most of its component features may get associated with very highly satisfactory non-anomalous readings in the third quartile of the reading range associated thereof. Under these situations, it is difficult for known machine learning techniques to correctly classify the signal that carries a fault condition and generate a necessary alert. Again, such situations also demand a mechanism to assess severity associated with the fault condition.
An exemplary illustration of the technical problem addressed by systems and methods of the present disclosure is as follows. A monitoring system within an Intensive Care Unit (ICU) generates an alert when a patient’s blood pressure (BP) rises above 140 or when the body temperature rises above 102ºF. The monitoring system also generates an alert if both the conditions i.e., high BP and high fever simultaneously happen for a patient. But the monitoring system does not assess the severity of the problem. For each of these alert situations, severities are quite different as implied by the BP level or by the body temperature.
i. BP = 143, Body Temp = 98.5º F
ii. BP = 143, Body Temp = 103 º F
iii. BP = 180, Body Temp = 98.5 º F
iv. BP = 155, Body Temp = 103 º F
v. BP = 133, Body Temp = 106 º F
vi. BP = 143, Body Temp = 104 º F
vii. BP = 200, Body Temp = 99.5 º F
In terms of severity levels, the alert cases numbered vii, v, and iii need to be highly prioritized than the rest. Next vi and iv needs to be ranked through severity levels. An understanding of this severity does not come through the identification of a failure signal alone. Even though the alert generation ensures desired attention of the respective authorities, it does not provide an indication of the extent of the problem that has occurred. In any diagnostic situation it is critical to reduce false positives and false negatives. Presently, in absence of a system and method to do the fault detection automatically and accurately, majority of the cases have to be dealt with manually by experts based on their experiences which may be costly and time consuming. Again, in absence of a measure of severity, prioritizing required responses cannot be done properly and may prove to be costly.
Referring now to the drawings, and more particularly to FIGS. 1 through 4, where similar reference characters denote corresponding features consistently throughout the figures, there are shown preferred embodiments and these embodiments are described in the context of the following exemplary system and method.
FIG.1 illustrates an exemplary block diagram of a system 100 for detection of faults in deceptive signals and computing severity thereof, in accordance with an embodiment of the present disclosure. In an embodiment, the system 100 includes one or more processors 104, communication interface device(s) or input/output (I/O) interface(s) 106, and one or more data storage devices or memory 102 operatively coupled to the one or more processors 104. The one or more processors 104 that are hardware processors can be implemented as one or more microprocessors, microcomputers, microcontrollers, digital signal processors, central processing units, state machines, graphics controllers, logic circuitries, and/or any devices that manipulate signals based on operational instructions. Among other capabilities, the processor(s) are configured to fetch and execute computer-readable instructions stored in the memory. In an embodiment, the system 100 can be implemented in a variety of computing systems, such as laptop computers, notebooks, hand-held devices, workstations, mainframe computers, servers, a network cloud and the like.
The I/O interface device(s) 106 can include a variety of software and hardware interfaces, for example, a web interface, a graphical user interface, and the like and can facilitate multiple communications within a wide variety of networks N/W and protocol types, including wired networks, for example, LAN, cable, etc., and wireless networks, such as WLAN, cellular, or satellite. In an embodiment, the I/O interface device(s) can include one or more ports for connecting a number of devices to one another or to another server.
The memory 102 may include any computer-readable medium known in the art including, for example, volatile memory, such as static random access memory (SRAM) and dynamic random access memory (DRAM), and/or non-volatile memory, such as read only memory (ROM), erasable programmable ROM, flash memories, hard disks, optical disks, and magnetic tapes. In an embodiment, one or more modules (not shown) of the system 100 can be stored in the memory 102.
In an embodiment, the system 100 includes one or more data storage devices or memory 102 operatively coupled to the one or more processors 104 and is configured to store instructions configured for execution of steps of the method 200 by the one or more processors 104.
FIG.2 is an exemplary flow diagram illustrating a computer implemented method for detection of faults in deceptive signals and computing severity thereof, in accordance with an embodiment of the present disclosure. The steps of the method 200 will now be explained in detail with reference to the components of the system 100 of FIG.1. Although process steps, method steps, techniques or the like may be described in a sequential order, such processes, methods and techniques may be configured to work in alternate orders. In other words, any sequence or order of steps that may be described does not necessarily indicate a requirement that the steps be performed in that order. The steps of processes described herein may be performed in any order practical. Further, some steps may be performed simultaneously.
Accordingly, in an embodiment of the present disclosure, the one or more processors 104 are configured to obtain, at step 202, boundary values (bi) for a plurality of signal features under consideration, the plurality of signal features being measured by a plurality of sensors. In the context of the present disclosure, the boundary values define limits for the plurality of signal features under consideration to classify associated signal features as belonging to a faulty group or a healthy group. Let there be ‘n’ number of sensors that send numeric readings as a signal from a subject being monitored. Each sensor monitors a particular feature that is considered business critical. Any failure occurring in a feature calls for immediate attention of authorities and qualifies the overall situation under a failure or a faulty or a specific clinical diagnostic category. While monitoring and diagnosing a problem condition, experts usually follow international, national or domain specific standards that specify tolerance limits on various diagnostic tests and conditions. The limits indicate a boundary of acceptance and are also indicative of whether an anomalous condition can be tolerated without an intervention. In other words, a specific limit may be treated as a boundary value between healthy and faulty or undesirable conditions specific to a problem feature. If fault conditions occur within a range, there may be two boundary points. Accordingly, each of the plurality of signal features is associated with one or more boundary values. For instance, a signal feature Blood Pressure is associated with two boundary values corresponding to systolic pressure and diastolic pressure.
In an embodiment, the step of obtaining boundary values comprises performing statistical analysis such as analyses disclosed in an article entitled “A Simple and Reliable Method for Clinical Assessment of Odor Thresholds” by Jörn Lötsch et al. or an article entitled “A Study on Threshold Selection for Multi-label Classification” by Rong-En Fan and Chih-Jen.
In an embodiment of the present disclosure, the one or more processors 104 are configured to obtain, at step 204, readings (si) from each of the plurality of sensors, wherein the obtained readings together represent a signal, and wherein a deviation from the boundary values towards the faulty group renders the associated readings as anomalous readings.
In an embodiment of the present disclosure, the one or more processors 104 are configured to identify, at step 206, the signal as a faulty deceptive signal when the signal is characterized by (i) a relatively small number of anomalous readings being obliterated by remaining non-anomalous readings from the obtained readings, wherein the relatively small number is decided empirically; and (ii) each of the remaining non-anomalous readings being above a third quartile of a reading range associated with each of the plurality of signal features. In an embodiment, the relatively small number is less than 20%.
In the type of problem situations mentioned earlier, it is observed that faulty situations are unidirectional, meaning if the obtained reading falls below a limit set by the standard then a problem is suspected and an intervention is necessary, for instance when hemoglobin level is monitored. Also, it can be the other way that when the obtained reading falls above a limit point a problem condition is suspected, for instance when Prostate Specific Antigen (PSA) level is monitored.
Thus,
n =
for i = 1 to n, let
s ?= {sensor readings}={s_i}
b ?= {boundary values}={b_i}
Sensors are designed to measure a feature that represents a problem condition.
A weight is introduced for each feature condition.
w ?= {feature weights}={w_i}
Features need to be appropriately scaled to ascertain a meaningful distance between two signals representing measurements from the feature conditions. This is required even when all features are equally important or equally critical to identify the problem condition.
In an embodiment of the present disclosure, the one or more processors 104 are configured to scale and weigh the readings (si), at step 208, to obtain a comparable distance between the pluralities of signal features comprising the signal.
In an embodiment, the step of scaling and weighing the readings is based on interquartile ranges (IQR) associated with the plurality of signal features. In the absence of domain knowledge about the range of each feature, a method to scale, learning from the available data, is described below for the purpose of distance measurement.
Let the portion of the population having a problem c ondition or a fault condition be denoted by F. The other portion, the healthy population is represented by H.
Combining interquartile ranges of ith feature that are extracted separately from each group F and H, constructing combined interquartile range Ri and weights wi as:
R_i=[(Q_3^((F_i ) )-Q_1^(?(F?_i)) )+(Q_3^((H_i))-Q_1^(?(H?_i)))]
and,
w_i=(Max_i |R_i |)/R_i
Weights are designed, such that w_i>0 and a failure condition is the one that falls short of boundary value or b_i=s_i. This indicates w_i (b_i-s_i )=0 corresponds to a problem condition.
For example in an examination, a score less than 30 may indicate a failing condition for a subject. If there are seven subjects then n = 7 and bi = 29 for each i = 1 to 7. For all i, Q3 > Q1, Ri > 0 and wi > 0. A score of 15 (si = 15) in a subject would indicate a failing condition along with wi (29 – 15) = 0.
Another example having a reverse direction of problem condition involves the PSA test from a blood sample that provides a score. A score greater than 4 generates an alert generally recommending prostate biopsy. In the PSA case, the bigger the score, the greater is the problem situation. So to remain consistent with the design that b_i=s_i corresponding to problem occurrence, it is needed to set bi = -4, and si = -10. Thus when a PSA score of 10 is obtained, the system remains consistent checking that: w_i ((-4) – (-10)) = 0 indicates a problem situation.
So the generalized convention, for the ith feature condition, is as follows:
{¦(w_i (b_i-s_i )<0, ok@otherwise, problem)¦

In an embodiment of the present disclosure, the one or more processors 104 are configured to compute, at step 210, a directional measure (d) representative of a distance of the signal from a classifying boundary for determining a level of severity of the faulty deceptive signal, wherein the classifying boundary is based on the obtained boundary values. The concept of distance and severity may be explained in terms of the following examples.
Example-1:
In a semester exam having seven different subjects and 40 being the pass mark for each subject, following are the two sets of test scores of two students who failed the semester.
Student-A = {57, 61, 23, 50, 54, 66, 71}
Student-B = {61, 52, 59, 68, 39, 37, 64}
It is seen that both the students have failed the overall exam. But there is a need to determine the severity of the failure.
Based on count, A failed in 1 subject and B failed in 2 subjects. In a multidimensional situation like the example of the test scores above, the count showing the number of failure points does not necessarily correlate with the severity of the problem.
Even though B failed in two subjects, he barely fell short of pass marks. Whereas A is pretty far away from passing the subject in which he failed.
Example-2:
Suppose, in a battery of medical tests from blood samples, following levels are checked against standard specifications (boundary values):
1. Glucose level > 200
2. Hemoglobin level < 13
3. Cholesterol level > 240
4. Infection level > 30
5. Calcium level < 8
Suppose an individual set of results for three patients A, B, C is as follows:
The result set corresponds to levels of {Glucose, Hemoglobin, Cholesterol, Infection, Calcium}.
A = {178, 5, 190, 18, 11}
B = {201, 14, 245, 20, 12}
C = {191, 15, 230, 78, 10}
Based on count, ‘B’ has two failure conditions whereas ‘A’ and ‘C’ has one condition each.
But the underlying medical conditions may imply:
• A has very severe condition with very low Hemoglobin, requiring urgent blood transfusion.
• C has infection that requires treatment and antibiotics.
• B has high Glucose and high Cholesterol requiring the person to make certain lifestyle changes to begin with.
Severity of a problem measures the extent of departure from the acceptable conditions. In the exemplary scenario illustrated above, it is clear that the ranking of conditions of A, B and C with respect to the severity of the medical problem needs to be as follows:
Severity of condition of A > Severity of condition of C > Severity of condition of B
To find a use of severity in context of the healthy people (or an acceptable condition), consider a similar test result set of two persons who are supposedly healthy.
E = {186, 14, 225, 19, 11}
F = {195, 14, 238, 28, 9}
Between E and F, the test results probably indicate the condition of E as better than the condition of F. Because F’s results appear more near to boundary values even though figuring within a safe side of the limits. A measure which would be capable to differentiate the severities of A, B and C should also at the same time be capable to determine that E and F are healthy as opposed to A, B and C. So the same severity measure when applied to healthy group should indicate, based on the test results, that:
i. Between E and F, severity of condition of F > severity of condition of E.
ii. Among all, severity of condition A > severity of condition C > severity of condition B > severity of condition F > severity of condition E. In comparison to the conditions of A, B and C, the conditions of E and F are in the opposite direction of the boundary.

Example-3:
Digital-Subscriber-Line (DSL) provides internet subscribers with high-speed internet access using copper lines. DSL relies on DSL-Access-Multiplexers to transmit digital or analog signals. Access Multiplexers can allocate bandwidth symmetrically or asymmetrically between downstream and upstream speeds. One of the major downsides of DSL is that speeds attenuate the farther away a subscriber is located from a telephone exchange or distribution point.
Network Access Control thresholds corresponding to Asymmetrical Digital Subscriber Line (ADSL) technology may appear as below:
(Target) Seeking Speed (Mbps) Loop Length (Km) Loop Resistance for 0.4 mm Copper (Ohm) Loop Resistance for 0.5 mm Copper (Ohm) Attainable Download (kbps) SNR
(Signal to Noise Ratio) Attenuation (dB)
2 < 2.5 < 743 < 457 > 2560 > 10 < 69
4 < 1.2 < 356 < 219 > 6016 > 10 < 56
8 < 1.0 < 297 < 183 > 10240 > 10 < 41
16 < 0.7 < 208 < 128 > 19500 > 7.5 < 25
There are many more feature variables of a network such as Actual bit rate for Down Speed and Up Speed (kbps), Electrical length (dB @ 1MHz), Longest bridged tap length from spectral files (m) etc. that are ignored for the sake of simplicity. Suppose, due to complaints from three households using an Internet Service, the provider identifies the location of the issue and a set of devices nearby to assess the situation through network access control parameters and finds that:
For household-1:
1. Seeking Speed 1.3 Mbps
2. Loop length 2.4 km
3. Loop resistance (0.4mm Copper wire) 670 Ohm
4. Attainable download 2550
5. SNR 12
6. Attenuation 58
For household-2:
1. Seeking Speed 1.1 Mbps
2. Loop length 2 km
3. Loop resistance (0.4mm Copper wire) 780 Ohm
4. Attainable download 2561
5. SNR 10
6. Attenuation 68
Based on ordinary count, ‘household-1’ has two threshold conditions outside the acceptable limit whereas ‘household-2’ has one such condition. Actual conditions might imply an onset of a highly severe problem in the network of ‘household-2’ due to very high loop resistance. The problem for the ‘household-1’ may be due to noise from a malfunctioning device and may be much less severe in nature than the other case. Severity of a problem tries to measure the extent of departure from the acceptable conditions. Here ranking of the problem conditions of the two exemplified households needs to be as follows:
Severity of condition of ‘household-2’ > Severity of condition of ‘household-1’.
The fact that readings approaching towards boundary/threshold values, even when it remains within healthy side of the limits, may imply an overall progression towards a problem condition. This aspect can be exploited to find a use of ‘severity’ in the context of a healthy network or a health condition to gauge its nearness to a problem condition. A notion of directional measure of distance comes into play to shape up such a measure of severity. A measure of distance from the boundary values can thus represent the severity in such clinical conditions.
In an embodiment of the present disclosure, the step of computing the directional measure (d) is based on a sum of the exponentiated differences between the signal features and corresponding boundary values as given below.
distance(from b ? of s ? )=d([b ? ],s ?)=log_(n+1)?[?_(i=1)^n¦?(n+1)?^(w_i (b_i-s_i)) ]
where
w_i=(Max_i |R_i |)/R_i
and
d: Rn ? R
It may be observed that,¦(max@i)(w_i (b_i-s_i )) dominates the log space.
With d representing a positive number that can be small if for most i’s, w_i (b_i- s_i )<0, and in consideration of the contributions of the dominating terms when w_i (b_i-s_i )>0, along with an integer n_0, (1=n_0=n) , the behavioral design of the directional measure is better linked through an equivalent expression as below:
log_((n+1))?[?_(i=1)^n¦?(n+1)?^(w_i (b_i-s_i)) ]=log_(n+1)?[n_0.?(n+1)?^(?max?_i w_i (b_i-s_i))+d]
Considering possible variations within a set of sensor readings, a semi-representative single sensor depiction of the corresponding variations in directional measure may be as represented in FIG.3 wherein the X-axis represents the difference (b_i-s_i ) and Y-axis represents the directional measure. An observation from FIG.3 is that as sensor readings keep shifting outside the acceptable limit, severity of the condition keeps increasing sharply, towards the positive Y axis. The severity measure for a problem condition becomes super sensitive to changes within the signal when any reading falls short of its boundary. In contrast, a healthy signal having acceptable readings gets easily separated out as the sharpness of the severity flattens out. The severity measure for a healthy case becomes insensitive or rather remains much understated in response to changes within the signal. In such cases, changes will only be visible in very high resolution or through higher decimal places of the fractional part of the distance measure. Due to such super sensitivity associated with a measure of severity in analyzing the faulty signals, it becomes easy, in accordance with the present disclosure, to identify a signal of fault occurrence wherein a majority of individual features contrastingly appear very highly satisfactory and may appear overall healthy from the outset.
In an embodiment, the system 100 of the present disclosure may develop alerts based on the following:

{¦(log_((n+1))??[?_(i=1)^n¦?(n+1)?^(w_i (b_i-s_i)) ]>0,alarm?@otherwise, ignore )¦
In an embodiment, a positive value for the directional measure (d) renders the signal being classified as belonging to the faulty group and a negative value for the directional measure (d) renders the signal being classified as belonging to the healthy group.
When a signal s ? hovers near its boundary b ? between the faulty and healthy group, chances of getting a fault condition approximately stays around 50%. With this notion in place, when signals change readings and move away from the boundary, the probability of fault occurrence (or of a problem condition) increases or decreases depending on the direction of the signal deviation from the boundary. After a certain point, in each direction, it becomes clear that either 100% or 0% chance of occurrence of the problem condition has set in. As an example, consider a simple case of assessing hearing-problems using sound signals of varying frequency levels. The range of human hearing is generally considered to be 20 Hz to 20 kHz. As humans get older, they hear less of higher frequencies, generally hearing limit gets reduced to 17 kHz and below and that is considered a normal occurrence. So if a patient cannot hear above 16 kHz, the person’s condition may appear slightly problematic or can be ignored. On the other hand, when a patient cannot hear sounds above 11 kHz, an expert can confidently conclude with about 100% certainty that the patient has a hearing problem. Also, for another patient who cannot hear sounds above 9 kHz, the expert can again confidently conclude with 100% certainty that the patient has a hearing problem. So the sensor readings beyond a certain point may become less significant compared to the fact that the suspected problem condition is already established with certainty. There is a chance aspect that encompasses the degree of certainty of fault occurrence considerations with respect to the boundary conditions of a problem condition. The degree aspect may vary across business domains having somewhat different interpretations of a specific chance percentage.
In an embodiment of the present disclosure, the one or more processors 104 are configured to convert, at step 212, the computed directional measure (d) to a probability of severity based on an associated business domain corresponding to the signal features under consideration using regression analysis. FIG.4 illustrates fitment of the computed directional measure (d) to the probability of severity based on an associated business domain, in accordance with an embodiment of the present disclosure. In an embodiment, the step of converting the computed directional measure (d) to a probability of severity comprises mapping the computed directional measure to a pre-defined probability of severity as discussed in the example of assessing hearing-problems, wherein an expert concludes with 100% probability that a patient who cannot hear sounds above 9 kHz or above 11kHz has a hearing problem. In an embodiment, the fitment of probability to the directional measure takes the shape of an S-curve, due to the observations explained in reference to FIG.3 above. The probability of severity provides an indication of the extent of the problem condition. Further, due to the extreme sensitivity associated with the measure of severity in analyzing the deceptive signals, the step 212 of the present disclosure establishes a way to identify signals carrying failure messages hidden or obscured by the fact that majority of constituting features manifest as highly satisfactory, obliterating the failing features.
Thus, in accordance with the present disclosure, systems and methods of the present disclosure facilitate identification of deceptive signals and computing an associated probability of severity. The systems and methods of the present disclosure can be adapted in applications dealing with diagnostic signals such as in the communication, media, manufacturing, healthcare industries and the like.
The written description describes the subject matter herein to enable any person skilled in the art to make and use the embodiments. The scope of the subject matter embodiments is defined by the claims and may include other modifications that occur to those skilled in the art. Such other modifications are intended to be within the scope of the claims if they have similar elements that do not differ from the literal language of the claims or if they include equivalent elements with insubstantial differences from the literal language of the claims.
It is to be understood that the scope of the protection is extended to such a program and in addition to a computer-readable means having a message therein; such computer-readable storage means contain program-code means for implementation of one or more steps of the method, when the program runs on a server or mobile device or any suitable programmable device. The hardware device can be any kind of device which can be programmed including e.g. any kind of computer like a server or a personal computer, or the like, or any combination thereof. The device may also include means which could be e.g. hardware means like e.g. an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or a combination of hardware and software means, e.g. an ASIC and an FPGA, or at least one microprocessor and at least one memory with software modules located therein. Thus, the means can include both hardware means and software means. The method embodiments described herein could be implemented in hardware and software. The device may also include software means. Alternatively, the embodiments may be implemented on different hardware devices, e.g. using a plurality of CPUs.
The embodiments herein can comprise hardware and software elements. The embodiments that are implemented in software include but are not limited to, firmware, resident software, microcode, etc. The functions performed by various modules described herein may be implemented in other modules or combinations of other modules. For the purposes of this description, a computer-usable or computer readable medium can be any apparatus that can comprise, store, communicate, propagate, or transport the program for use by or in connection with the instruction execution system, apparatus, or device.
The illustrated steps are set out to explain the exemplary embodiments shown, and it should be anticipated that ongoing technological development will change the manner in which particular functions are performed. These examples are presented herein for purposes of illustration, and not limitation. Further, the boundaries of the functional building blocks have been arbitrarily defined herein for the convenience of the description. Alternative boundaries can be defined so long as the specified functions and relationships thereof are appropriately performed. Alternatives (including equivalents, extensions, variations, deviations, etc., of those described herein) will be apparent to persons skilled in the relevant art(s) based on the teachings contained herein. Such alternatives fall within the scope and spirit of the disclosed embodiments. Also, the words “comprising,” “having,” “containing,” and “including,” and other similar forms are intended to be equivalent in meaning and be open ended in that an item or items following any one of these words is not meant to be an exhaustive listing of such item or items, or meant to be limited to only the listed item or items. It must also be noted that as used herein and in the appended claims, the singular forms “a,” “an,” and “the” include plural references unless the context clearly dictates otherwise.
Furthermore, one or more computer-readable storage media may be utilized in implementing embodiments consistent with the present disclosure. A computer-readable storage medium refers to any type of physical memory on which information or data readable by a processor may be stored. Thus, a computer-readable storage medium may store instructions for execution by one or more processors, including instructions for causing the processor(s) to perform steps or stages consistent with the embodiments described herein. The term “computer-readable medium” should be understood to include tangible items and exclude carrier waves and transient signals, i.e., be non-transitory. Examples include random access memory (RAM), read-only memory (ROM), volatile memory, nonvolatile memory, hard drives, CD ROMs, DVDs, flash drives, disks, and any other known physical storage media.
It is intended that the disclosure and examples be considered as exemplary only, with a true scope and spirit of disclosed embodiments being indicated by the following claims.