M. M. Sufyan Beg

Soft Computing Techniques for Rank Aggregation on the World Wide Web

Rank aggregation is the problem of generating a near-“consensus” ranking for a given set of rankings. When applied to the web, this finds applications in meta-searching, search engine comparison, spam fighting and word association techniques. The rank aggregation obtained by optimizing the Spearman footrule distance is called footrule optimal aggregation (FOA), and it also satisfies the Condorcet property. We find in literature a polynomial time algorithm to compute FOA for full lists. However, when collating the results of the search engines, the lists are almost invariably always the partial ones, as different search engines usually return non-overlapping lists of documents. The FOA for partial lists, however, is NP-hard. This NP-hard nature of partial footrule optimal aggregation problem (PFOA) motivates us to apply genetic algorithm (GA) for the PFOA problem. The GA based technique may take long to compute, but we propose to decide upon the number of generations of GA based on the time limit allowed by the user. We have also considered some “positional” methods, as they are linear in complexity. A classical positional method is the Bordas method. Since, fuzzy logic has been extensively studied in literature for arriving at consensus in group decision making, the adoption of some fuzzy techniques is also being investigated here for getting an improvement over the Bordas method. We have not only adopted and compared the classical fuzzy rank ordering techniques for web applications, but also proposed three novel techniques that outshine the existing techniques.

Elements of Sketching with Words

In modern day crimes and terrorism, it has become imperative to identify by features the criminals who are involved and who have caused such a disaster. The current work is a basic step towards such an important identification process. In this direction, we begin our work with the identification of fuzzy geometric shapes, which resemble with the actual geometric shapes fuzzily. Zadeh proposed Computing with Words (CW) which deals with perceptions, where the perceptions are mostly fuzzy, while the measurements are always found to be crisp. On the similar lines, Sketching with Words (SW) is a technique in which the objects of sketching are the perceptions described in uncertain, vague, imprecise words and prepositions described in Natural Language (NL). We propose SW here to represent geometric figures using membership functions to estimate f-geometry.

Face sketch recognition using sketching with words

The face sketch of the criminal may be one of the crucial evidence in catching the criminal. Face sketch is drawn by the sketch expert on the basis of onlooker’s statement, which is about different human face parts like forehead, eyes, nose, and chin etc. These statements are full of uncertainties e.g. ‘His eyes were not fairly small’. Since the precise interpretation of these natural language statements is a very difficult task. So we need a system that can convert imprecise face description, into a complete face. Therefore we have applied the sketching with words (SWW) technique to design a system that can simulate a face sketch expert. SWW is a methodology in which the objects of computation are fuzzy geometric objects e.g. fuzzy line, fuzzy circle, fuzzy triangle, and fuzzy parallel. These fuzzy objects (f-objects) are formalized by fuzzy geometry (f-geometry) of Zadeh. SWW is inspired by computing with words and fuzzy geometry. Since the onlooker has to granulate face into granule label. Hence the concept of fuzzy granule has applied for face recognition. Different types of face have generated after applying ‘fairly’ and ‘very’ linguistic hedges on face components.

Automatic Performance Evaluation of Web Search Systems using Rough Set based Rank Aggregation

Web searching is such an activity that its importance can just not be ignored in the current scenario. Since there are a large number of publicly accessible search engines, shich differ in their indexing algorithms and hence the search results, the evaluation of search engines performance is needed to determine which one is the best. The human intelligence may be used to measure the search engine effectiveness. But, a subjective evaluation done on the basis of user-feedback is costly in terms of the time required. Therefore, it is also not scalable. So, there is a need of an automatic evaluation method. In this paper, we present the architecture of an automatic Web search evaluation system that combines the different evaluation techniques using a Rough Set based Rank aggregation technique. The rough set based rank aggregation models the user’s feedback based rank aggregation. In the rough set based aggregation technique, the ranking rules are learnt on the basis of the user feedback in the training data sets. The learned rules are then used to estimate the overall ranking for the other data sets, for which user feedback is not available. We show our experimental results pertaining to seven public search engines.

PNL-Enhanced Restricted Domain Question Answering System

The concept of PNL (Precisiated Natural Language) has been proposed by Zadeh for computation with perceptions and some problems described in natural language. We describe a design for restricted domain question answering systems enhanced by PNL-based reasoning. For a subset of a knowledge corpus (e.g. critical or frequently-asked topics) where fuzzy set definitions of vague terms are provided, more precise answers can be computed via protoformal deduction. Nested structure in the system design also enables processing of natural language statements that are not PNL protoforms using phrase-based deduction and concept matching to generate the most relevant facts for a query. If deduction results yield low confidence factor, standard search engine provides a baseline response (relevant paragraphs based on keyword matches). Our design principles aim for flexible, domain independent capability and minimize human input to provision of semantic clues and background knowledge during design or application set-up.

A new computational fuzzy time series model to forecast number of outpatient visits

Forecasting number of outpatient visits is pre-eminent for patient planning, medical resource utilization and overall management of health care system to a certain extent. Aim of forecasting the outpatient visits can also be seemed in terms of individual care. In addition, accurate prediction of outpatient visits in hospitals can play a significant role in health insurance plans and for deciding reimbursement system. As such, main challenge in healthcare simulation is to produce a realistic model that must utilize efficient techniques for managing complex time series data and should be capable of generating forecasted value with almost negligible error. We proposed forecasting model based on fuzzy time series that rectifies the existing imperfections and overcome the drawbacks of previous approaches. Novice concept introduced to eliminate the inadequacies by way of defining the universe of discourse on historical data. Model also endeavors to pontificate the issue of improving forecasting accuracy through the new idea of event discretization function. This was quite encouraging as it highlights the impact of trend & seasonal components by yielding dynamic change of values from time t to t+1. This fuzzy computing time series model is designed by joint consideration of three key points (1) Event discretization of time series data (2) Frequency density based partitioning (3) Creation of Fuzzy logical relationships in optimized way. Subsequently, performance of the proposed model is demonstrated and compared with some of the pre-existing forecasting methods on same outpatient data. In general, findings of the study are interesting and superior in terms of least Average Forecasting Error Rate (AFER) and Mean Square Error (MSE) values.

Fuzzy time series model to forecast rice production

Crop production is considered as one of the real world complex problem due to its non-deterministic nature and uncertain behavior. Particularly, forecasting of rice production for a lead year is pre-eminent for crop planning, agro based resource utilization and overall management of rice production. As such, main challenge in rice production forecasting is to generate realistic method that must be capable for handling complex time series data and generating forecasting with almost negligible error. The objective of present work is to design & implement such a competent fuzzy time series model for forecasting of rice production. We have proposed forecasting model based on fuzzy time series that highlights the impact of trend & seasonal components by yielding dynamic change of values from time t to t+1. The aim of using fuzzy time series is to deal with forecasting under the fuzzy environment that contains uncertainty, vagueness and imprecision. This method assigns importance to fuzzy intervals on the basis of frequency of number of time series data. Subsequently, computed fuzzy logical relations are used for analysis of time series rather than random and non-random functions as in case of usual time series analysis. Performance of the proposed model is demonstrated and compared with few pre-existing forecasting methods on rice production. To prove robustness and accuracy of the presented model, analysis is performed on forecasting of enrollment data of university of Alabama.

Modified rough set based aggregation for effective evaluation of web search systems

Rank Aggregation is the problem of generating a single consensus ranking for a given set of rankings. Rough set based Rank aggregation is a user feedback based technique for rank aggregation, which learns ranking rules using rough set theory. In this paper, we discuss an improved version of the Rough set based Rank aggregation technique, which is more suitable for aggregation of different Web search evaluation techniques. For learning the ranking rules, we obtain the implicit user feedback to the search results returned by a search engine in response to a set of fifteen queries and mine the ranking rules using rough set theory. In the modified rough set based rank aggregation technique, we incorporate the confidence of the rules in predicting a class for a given set of data. That means, we do not say surely that the record belongs to a particular class according to a particular rule. Instead, we associate a score variable to the predicted class of the record, where the value of the variable is equal to the confidence measure of the rule. We validate the mined ranking rules by comparing the predicted user feedback based ranking with the actual user feedback based ranking. We apply the ranking rules to another set of thirty seven queries for aggregating different rankings of search results obtained on the basis of different evaluation techniques. We show our experimental results pertaining to seven public search engines.

Fuzzy Identification of Geometric Shapes

The identification of criminals with sketches can no longer sustain using conventional image processing techniques. Since, it behaves mechanistically, that is, a system which behaves as per given set of rules. We propose a humanistic system for identification of sketches of criminals. Certainly, one must be looking forward for a novel approach, which identifies similarity between a photographic image and a transformed fuzzy image i.e., a sketched image. The transformation on images could be anyone among rotation, reflection, translation, scaling or shearing. In this regard, our approach identifies fuzzy geometric shapes, like humans identify any imprecise shape with their cognition. Such fuzzy shapes cannot be left unidentified under crucial conditions. We begin with estimation of f-validity and then the f-similarity for f-geometric objects, which are considered as basics for developing a humanistic identification system. We implement OWA operators for computing f-similarity in fuzzy geometric shapes. Moreover, the results are found to be justified with the extent of fuzziness.

Enhanced accuracy of fuzzy time series predictor using genetic algorithm

Accuracy is one of the most important aspects in the domain of forecasting. It is very difficult to improve accuracy of prediction system where prediction is based only on large historical values and accuracy is important for each predicted value along with the whole system. The main objective of this research is to optimize dominant factors of fuzzy time series predictor (FTSP) using genetic algorithm (GA) and further to improve prediction accuracy for each time series variable along with whole system. This is obtained by (a) generating wide range of parameters for membership function at time t on the basis of their base value (b) subset of population generated at time t is used for fitness checking. Additionally, GA complexity is also reduced by utilizing rate of change of time series data to cut down the bit size of chromosome. It can be observed from comparative study that use of GA considerably reduced mean square error (MSE) and average forecasting error rate (AFER).