Facebook Twitter Digg Reddit Linkedin StumbleUpon E-mail

Research Article

An Adaptive Approach for Real Time Advertising System based on User Location and Interests

Journal of Software Engineering: Volume 11 (2): 160-171, 2017

Mohammed A. Abuhelaleh, Ayman Al Dmour and Tahir Aletewi

Abstract

Background: Due to the wide spread of the smart phones around the world, a special demand occurs in the mobile marketing. Real time advertising is still challenging due to its hard requirements: Tracking of customer location, knowledge of customer interests, knowledge of advertisers data and the ability to efficiently provide the customer with available advertiser’s data securely. Materials and Methods: This study introduces a comprehensive work for designing, developing and evaluation of a mobile platform for real time advertising system. This system is built by adopting well-known algorithms for each requirement concerning real time advertising system. To investigate the performance of the developed system, several experiments are conducted using a simulation program built specially for this system to measure two main factors: The missing advertisements and the delay time. Results: The simulation results show that the performance of the proposed system is very promising and revealing the importance and possibility of developing adaptive real time advertising system based on customer’s location and interest. The proposed system considered as a promising and a stable media for advertisement according to the missing advertisements and the time required for the advertisement to be received by the customer. Conclusion: This study introduce a novel system, which is built to satisfy user interests for advertisements based on his location. The system built using three subsystems to ensure its ability to produce data in real time and to handle huge traffic of subscribers and advertiser. It has also the ability to integrate different systems into it to satisfy any future requirements.

How to cite this article:

Copyright
© 2017. This is an open access article distributed under the terms of the creative commons attribution License, which permits unrestricted use, distribution and reproduction in any medium, provided the original author and source are credited.

Mohammed A. Abuhelaleh, Ayman Al Dmour and Tahir Aletewi, 2017. An Adaptive Approach for Real Time Advertising System based on User Location and Interests. Journal of Software Engineering, 11: 160-171.

DOI: 10.3923/jse.2017.160.171

URL: https://scialert.net/abstract/?doi=jse.2017.160.171

INTRODUCTION

A new competition field appears recently between companies to draw the attention of the customers to their products and services. This field produced from the wide spread of the smart phones devices around the world, with the integrated technologies they have. These technologies provide the customers with variety of choices not only in their area, but also around the world by a click. The appearance of this field of competition made the traditional advertisement systems limited and insufficient.

Real time advertising requires prior knowledge of user location and interests. In addition, it requires scalable subsystems with high performance to handle advertisements retrieval. Some researchers proposed techniques to retrieve accurate user location using the Geographical Positioning System (GPS) and other devices. Other researchers proposed techniques to analyze user behavior in order to put the hand on his interests. User privacy methods over mobile applications are proposed by some other researchers in order to guarantee the security of user data against attacks. On the other side, many algorithms have been recently proposed in order to retrieve big data (i.e., videos, images and texts) via internet. Each of these areas is a rich field for study.

This study invests the opportunities of well-known algorithms from these areas in the real time advertising system based on user location and interest (Fig. 1).

Existing algorithms and techniques: Some of the most recent proposed algorithms and techniques for each part of the system requirements are discussed to frame the available choices that can be used to build such systems (Fig. 1).

User location: The GPS device communicate with satellites to determine geographical locations for static and dynamic objects. This section discusses some techniques and algorithms proposed to get accurate data regarding object location from GPS devices.

Taylor and Sennott1 designed an algorithm called A-GPS referred to assisted GPS. This algorithm invests the assistance data (i.e., time stamp, navigation data, ephemeris information and the nearby station location), which is produced from GPS signal to enhance performance. Akopian and Syrjarinne2 followed the main structure of A-GPS and proposed an algorithm to analyze the GPS signal without decoding in order to produce more efficiency regarding the time required.

On the other side, Misra and Enge3 applied the concepts of A-GPS algorithms to decode the GPS signals and to evaluate the performance measurement of that signal in order to produce more data from that signal without affecting the performance of the main algorithm. Lin and Tsui4 solved the problem of the weak GPS signal by designing a special receiver that can handle and analyze GPS weak signals.

Fig. 1:Real time advertisement system requirements

Agarwal et al.5 did a survey on some algorithms to provide an idea of the A-GPS algorithm and other related algorithms drawbacks and the proposed solutions for these drawbacks.

User behavior: The behavior of the user may indicate his interests and concerns. Some of the techniques and algorithms concerning the study of user behavior are discussed. User behavior may produce qualitative information regarding his interests. The user-systems deal with the users according to these interests. The researchers divide the algorithms in this field mainly into two categories: Content-based recommender systems and context-aware recommender systems. Content-based recommender systems focus on description of the available system item/service and the user profile. On the other side, context-aware recommender systems focus on the contextual information that can be gathered from user location in order to analyze user behavior and interests.

Blanco-Fernadeze et al.6 proposed an algorithm based on a technology relates to semantic web reasoning in order to clarify the relationship between the item/service and the user preferences. Meng and Chen7 proposed Bayesian theory to filter user’s characteristics and item’s/service’s characteristics to get the best match. Li et al.8 proposed an acquisition method to deal with invisible data in order to build user model.

The previously discussed algorithms deal with current user behavior to analyze his interests. Other algorithms invested the user’s past behavior to build a complete knowledge of the current user interests. Lu9 proposed a filtering system to study purchases history of the users neighbors in order to get some expectations of user interests. Deng et al.10 proposed a clustering algorithm to apply a classification on the items/services which then can be used in other’s recommender systems.

Privacy: Some algorithms regarding user privacy over the internet. The main issue that concerns the users who are using the internet is to know how their data should be handled via the internet communication with the existence of hackers and spammers.

Guha et al.11 proposed a system to ensure user privacy. The profiling part processed at the client side rather than the broker side. The advertisement communication channel is protected by a special proxy to ensure communication securely. Client side is also monitored by a part of the system called reference monitor to ensure that the client obeys the system protocol. The idea of their solution is isolate the user out of the broker to ensure user security.

Toubiana et al.12 proposed a cryptography system to ensure the security of the user via advertisement system. They rely on the user side to manage the profiling and targeting step in the advertisement system. This prevents the advertiser from knowing the specific advertisement the user had choose in order to protect user privacy.

Guha et al.13 also proposed a solution called Koi for preserving privacy, which is based on the un-likability and an exchange protocol between two non-clouding servers and the user. This provides general solution that may satisfy different application types.

Big data: Transmitting of huge data over the internet is an existing challenge due to the limitations of transmitting channels, the storage limitation and the devices limitations. This section discusses some algorithms and techniques that have been proposed to overcome these challenges. Multimedia retrieval categorized mainly into three techniques: Multimedia indexing, multimedia compression and multimedia transmission (i.e., retrieval). These categories may involve specific type of multimedia (i.e., videos, text and images).

For multimedia indexing, a text-based video indexing was proposed by Li and Doermann14 to produce approximate matching of the context by expanding the metadata of the query using a matching method called glimpse matching. Lim et al.15 proposed a simple and fast methodology based on assumption that text background should include lower intensity than the text itself. This assumption makes it easy to distinguish the text from the background depending on a counted threshold. However, it does not work properly on a colored text, where it is hard to distinguish a colored background from a colored text.

For multimedia compression, ITU-T team developed H.261 and H.263 standards for video codec16,17. The ISO/IEC team developed MPEG-1 and MPEG-4 standards for video codec18,19. Then both teams worked together and they developed H.262/MPEG-2 and H264/MPEG-4 video codec standards20,21. The recent technique developed by Bross et al.22 was HEVC standard for high efficiency video coding.

For multimedia transmission, Yang et al.23 proposed a methodology to choose the best compression parameters for image compression in order to reduce the transmission energy consumption. They used a joint channel coding technique to transmit compressed images via bandlimited channel based on OFDM turbo coded. Tarokh et al.24 proposed transmit diversity methodology called a Space Time Block Coding (STBC) to enhance the network capacity and error performance for wireless systems in order to transmit more data via communication channels. Lin et al.25 improved the performance of STBC by combining it with adaptive beam forming in a correlated channel.

Advertisement systems: Evans et al.26 proposed iMAS system. The iMAS presents an advertisement system based on user location and interests. The user location updated frequently in order to get the available advertisers around that location. The user interests gathered as inputs from users via the user interface. Advertisement material then transmitted to the user when matching is occurs. The system has iMAS scheduler server, which is responsible of the whole processes. This makes the system incapable of handling big number of users in real time. In addition, the system design does not support the integration of other systems into it. This includes: Multimedia retrieval, wireless transmission and security.

Babu et al.27 applied the iMAS system as a mobile application. This application stores all matched advertisements for the user to be browsed on demand. Patil et al.28 proposed location based advertisement system. The system depends on user location to provide the advertisement material regardless of his interests. The system also involves administrator as a third party between the user and the system, which reduces the level of security and eliminates the user privacy. The system design also does not support the ability to integrate other solutions into it.

Our proposed system is designed to satisfy user interests depending on his location in real time with the consideration of user privacy. It has the ability to manage huge number of users in real time. Also, it has the ability to integrate different solutions for multimedia retrieval, wireless transmission, user behavior and internet security.

MATERIALS AND METHODS

Proposed system: The proposed system evolves many tasks to accomplish its function. Customer registers itself in the system and provides its profile and interests. The system keeps tracking the customer behavior securely to update his profile and interest accordingly. Advertiser registers himself in the system and provides his data and advertisement material. All data should be privately stored and retrieved when necessary using simple algorithms that keep the data secure, retrieve the data effectively and produce low overhead. All data should be stored securely in the cloud. The system then continuously get the customer-location updates as he walk or drive around advertiser’s area. According to customer location, interests and available advertisement material, the customer receives the data from the system using an algorithm that is able to compress and decompress the data in effective manner to be able to handle big data transmitting.

The RTAS system consists of three main sub systems which are: Tracking subsystem (Track the subscriber), notification subsystem (Notify the subscriber for new advertisements in the area) and retrieval subsystem (retrieve advertisements from the database). Figure 2 shows the three subsystems and their interactions.

The advertiser web interface enables the advertiser agency to manage his advertisements via the advertisement retrieval subsystem. These advertisements are stored into a database while the notification subsystem is updated with the advertisements preferences. The area is divided by equal zones area with a specific size for each zone, which is controlled by the zone manager. The notification subsystem updates the related zone manager with the new changes.

Fig. 2:RTAS subsystems and their interaction

Fig. 3:RTAS use case

The subscriber advertisements mobile application informs the notification subsystem with its preferences for preferred advertisements. The notification subsystem registers the information for later use. While the subscriber travels within the area, the tracking subsystem receives the subscriber location to check if it is within the advertisement range. If the advertisement is within the advertisement range, the tracking subsystem informs the notification subsystem in order to notify the subscriber with this advertisement. The subscriber then stores this notification for later use or view the advertisement using the advertisement retrieval sub system.

For user location retrieval, the proposed system adopts the improved A-GPS algorithm proposed by Akopian and Syrjarinne2 analyze the GPS signal without decoding in order to produce more efficiency regarding the time required.

The user interests determined by the subscriber at registration time and it can be updated occasionally. In addition, an integrated system to analyze user behavior is applied using the Bayesian theory proposed by Meng and Chen7 to filter user’s characteristics and the advertisements characteristics to get the best match.

To preserve the privacy between system parties, Koi solution proposed by Guha et al.13 is used. Koi is follows the un-likability and an exchange protocol between two non-clouding servers and the user.

The system applies HEVC standard for high efficiency video coding which is proposed by Bross et al.22. In addition, it applies a text-based video indexing which was proposed by Li and Doermann14 to produce approximate matching of the context by expanding the metadata of the query using a matching method called glimpse matching.

For multimedia transmission, the improved STBC methodology proposed by Lin et al.25 is applied to the proposed system to enhance the network capacity and error performance for wireless systems in order to transmit more data via communication channels.

System analysis: The analysis of the system involves: Defining user requirements, demonstrating system design basis and determining system testing requirements.

Use case diagram: The use case model provides a clear understanding of how the users of the systems (Actors) interact with the system. Each use case represents a complete scenario that starts with an event (from the actor) and changes the state of the system. Figure 3 shows RTAS use case model.

Fig. 4:RTAS class diagram

The system has two main actors:

•  Customer (subscriber): Who is interested in receiving advertisements
•  Advertiser: Who provides the advertisements

The secondary two actors:

•  User: Who is used for security purposes and can be either a customer or an advertiser
•  GPS provider: Who change the coordinates of the customer location and must be attached to the customer

The system has six main scenarios that describe the system:

•  Authenticate and register: Describe the security authentication measures
•  Update preferred advertisements and update location: Describe what categories and attributes that searches query should contain and when to receive an advertisement
Post advertisement: Describes what should be done to acquire and store an advertisement
Update advertisement: Describes what should be done to update and store an advertisement

A secondary scenario:

•  Pay fees: Describes what to do to ensure that a payment is completed

Class diagram: A class diagram model shows the classes that compose the system and their relations. Figure 4 shows RTAS class diagram.

Activity diagram: Activity diagram model shows activates (consist of operations) that the system carries and the time line, in which it is carried. In other words, it represents the system operations dynamically. Figure 5 shows the main activity diagram.

From a client view, the activity starts by checking the client authentication type. If the client is an advertiser, the flow goes to update management activity.

Fig. 5:RTAS main activity diagram

Fig. 6:RTAS location-change state diagram

If the client is a subscriber, then the system carries the advertisement to online list operation, which activates this subscriber as a node that is ready to accept advertisements. A couple of operations can be carried concurrently as shown in the figure. The subscriber can change his preferences while receiving notifications or he might sign off from the system.

State-chart diagram: State chart model represents the different states that the system can demonstrate according to some external or internal events. The RTAS state-chart for location change is presented in Fig. 6. Figure 7 shows the state of advertiser object. Figure 8 demonstrate the subscriber object state.

Fig. 7:RTAS advertiser state diagram

System design: The design of the proposed system depending on the system analysis presented in the previous. To clarify the design of the proposed system, component diagram is illustrated in Fig. 9. The system is divided mainly into two sides: Client side and server side.

The client side demonstrates the user side which involves the advertiser and the subscriber. In this part, user provides the system with all required information needed for advertisement process. Advertiser provides (using web application interface): Advertiser profile, locations and advertisements material. Subscriber provides (using mobile application interface): Subscriber profile, interests and location. Some information provided at system setup time and other information provided during system run.

At the server side, data is processed by different subsystems. The tracking location subsystem keeps tracking of subscriber location changes and informs the Ads notification subsystem with these changes for each online user.

Fig. 8:RTAS subscriber state diagram

The Ads notification subsystem stores the updates in the database frequently and with users related interests to specific advertisements in a specific zone. It then informs the Ads Retrieval subsystem with list of users and related advertisements information. The Ads retrieval subsystem receives the advertisement information from the advertiser and updates the database with this information. In addition, it retrieves the advertisement material from the database according to the information received from Ads notification subsystem and it sends it to the selected users.

RESULTS

The experimental results illustrated from running the proposed system in a virtual environment in order to analyze the performance of the system under different circumstances. A special simulation is built to demonstrate the system and to have the ability to apply different criteria for each run. The result focused on two main factors: The missing advertisements (i.e., the advertisements sent to the user but did not received by that user) and the delay time (the time duration from entering the zone area by the subscriber until receiving the advertisement). These factors are analyzed in different environments affected by three elements: Advertisers range, subscriber movement speed and zones sizes.

The simulation runs for 30 times to get the average results for each factor.

Fig. 9: Component diagram

Fig. 10(a-b): Advertisement range affection on (a) Missing advertisements and (b) Time delay

Table 1:Experiment setup values

The initial values for all runs were adjusted as the following: Simulation area size is set to 60000 m2 (average city size) with zone size that is set to 6000 m2, number of users is set to 1000 users, number of advertisements is set to 1000, number of preferred advertisements is set to a random number ranges 1-5. Maximum number of advertisements per zone is set to be a random number ranges 0-1000, but tends normally to try and distribute the advertisements uniformly, advertisement types per zone is set to be a random number ranges 1-5. Table 1 shows the summery of the experiment setup values.

First runs aimed to analyze the effect of the advertisement range on the number of missing advertisements (the advertisements that are not received by the subscriber because he left the zone area before receiving them) and the delay time (the period taken since the moment the customer entered the advertisement range to the actual time the advertisement reaches the customer phone). The number of users for this simulation is set to a 1000 users and the zone size is set to 6×6 m.

Figure 10a shows the result of the first round. The x-axis refers to the advertisement radius by meters starting from 200-1000 m range, while y-axis refers to the percentage of missing advertisements. The result shows that the number of missing advertisements slightly increases by increasing the advertisement range and it reaches 2% missing advertisements in 1000 m range.

Figure 10b show the results of the second round. The x-axis refers to the advertisement radius by meters starting from 200-1000 m range, while y-axis refers to the delay time in millisecond starting from 0-1500 msec. The result shows that the delay time is decreased significantly when the advertisement range reached 300 m. The delay time then remains stable after reaching 500 msec.

Second runs aimed to analyze the effect of the user speed on the number of missing advertisements and the delay time. The advertisement range for this simulation is set to a random number between 100-1000 m and the zone size is set to 6×6 m. Figure 11 illustrates the results produced by these runs. Figure 11a shows the results of the first round. The x-axis refers to the user speed in km h–1 starting from 5000-55000 m h–1 (within city speed limits for most cities), while y-axis refers to the percentage of missing advertisements.

The result shows that the missing advertisements ratio does not affected by the speed of the user. Figure 11b shows the results of the second round. The x-axis refers to the user speed in km h–1 starting from 5000-55000 m h–1, while y-axis refers to the delay time in millisecond starting from 0-3000 msec. The result shows that the delay time is significantly decreasing by increasing the user speed, where the delay time is close to zero when the user speed reaches 55000 m h–1.

Third runs aimed to analyze the effect of the zone size on the number of missing advertisements and the delay time. The advertisement range for this simulation is set to a random number between 100-1000 m and the user speed is set to 30 km. Figure 12 illustrates the results produced by these runs.

Figure 12a shows the results of the first round.

Fig. 11(a-b): User speed affection on (a) Missing advertisements and (b) Time delay

The x-axis refers to a square zone length starting from 1 km (i.e., zone area size of 1000 m2) to 11000 m (i.e., zone area size of 121000 m2) for each zone, while the y-axis the percentage of missing advertisements. The result shows that missing advertisement ratio is decreasing significantly by decreasing the zone size. The missing advertisement ratio reached a value closes to zero when the zone size exceed 9000 m (i.e., zone area size of 81000 m2).

Figure 12b shows the results of the second round. The x-axis refers to zone radius starting from 1 km (i.e., zone area size of 1000 m2) to 11000 m (i.e., zone area size of 121000 m2) for each zone, while y-axis refers to the delay time in millisecond starting from 0-600 msec. The result shows that the delay time does not significantly affected by zone size. The delay time ranges between 400 and 500 msec.

DISCUSSION

The system design discussed before and the simulation results shows the advantages of the proposed system comparing to other systems discussed previously.

Fig. 12(a-b): Zone size affection on (a) Missing advertisements and (b) Time delay

The iMAS system design proposed by Evans et al.26 and the modified iMAS system proposed by Babu et al.27 have a centralization problem. All communication traffic are centered in one system called scheduler server. This produces more cost regarding time, which results in losing the nature of real time concept in high traffic environment. In this proposal, data traffic is handled by different subsystems to ensure the ability to handle the high traffic produced from large number of users and advertisers.

The proposed system in this study handled the user privacy, which is not mentioned in the previous proposed systems. The privacy occurred from the way the systemhandle user data during system lifecycle. The data is encrypted for the subscriber and the advertiser before is being transmitted in network using Koi system proposed by Guha et al.13. The system also has no third party involved in the system. This produced more privacy regarding the data stored in the system. This privacy not considered by the system proposed by Patil et al.28.

The proposed system also applied most know techniques to handle big data processing and transmission to reduce the time and the space needed for the system processing. This ability is not available in the other proposed Real Advertising systems.

CONCLUSION

This study proposed an adaptive approach for real time advertising system based on user interests and location. The system communicates between the customers and the advertisers as an advertisement media in a real time fashion depending on customer location and interests. The advertisement area divided into zones to handle customer movement momently. The results show that the system provides a stable media for advertisement according to the missing advertisements and the time required for the advertisement to be received by the customer. The system is an adaptive due to the continuous updating of customer location and interests to provide suitable service on time.

References

Agarwal, N., J. Basch, P. Beckmann, P. Bharti and S. Bloebaum et al., 2002. Algorithms for GPS operation indoors and downtown. GPS Solutions, 6: 149-160.
CrossRefDirect Link

Akopian, D. and J. Syrjarinne, 2002. A network aided iterated LS method for GPS positioning and time recovery without navigation message decoding. Proceedings of the Position Location and Navigation Symposium, April 11-15, 2002, Las Vegas, NE., pp: 77-84.

Babu, B.R.P., B.L. Jaya Kumar and V.R. Bharatesh, 2014. An intelligent android Mobile based real time ads tracking system. Int. J. Adv. Res. Comput. Commun. Eng., 3: 7287-7292.
Direct Link

Blanco-Fernandez, Y., J.J. Pazos-Arias, A. Gil-Solla, M. Ramos-Cabrer and M. Lopez-Nores, 2008. Providing entertainment by content-based filtering and semantic reasoning in intelligent recommender systems. IEEE Trans. Consumer Electr., 54: 727-735.
CrossRefDirect Link

Bross, B., W.J. Han, G.J. Sullivan, J.R. Ohm and T. Wiegand, 2012. High Efficiency Video Coding (HEVC) text specification draft 9. Document JCTVC-K1003, ITU-T/ISO/IEC Joint Collaborative Team on Video Coding (JCT-VC), October 2012, China.

Deng, A.L., Z.Y. Zuo and Y.Y. Zhu, 2004. Collaborative filtering recommendation algorithm based on item clustering. Mini-Micro Syst., 25: 1665-1670.

Evans, C., P. Moore and A. Thomas, 2012. An intelligent mobile advertising system (iMAS): location-based advertising to individuals and business. Proceedings of the 6th International Conference on Complex, Intelligent and Software Intensive Systems, July 4-6, 2012, Palermo, pp: 959-964.

Guha, S., B. Cheng and P. Francis, 2001. Privad: Practical privacy in online advertising. Microsoft Research India. http://research.microsoft.com/en-us/um/people/saikat/paper-nsdi11-guha-privad.html.

Guha, S., M. Jain and V.N. Padmanabhan, 2012. Koi: A locationprivacy platform for smartphone apps. Proceedings of the 9th USENIX Conference on Networked Systems Design and Implementation, April 25-27, 2012, San Jose, CA, USA. pp: 14-.

ISO/IEC., 1993. Information technology-Coding of moving pictures and associated audio for digital storage media at up to about 1,5 Mbit/s-Part 2: Video. International Standard ISO/IEC 11172-2:1993, ISO/IEC, Switzerland, pp: 1-6.

ISO/IEC., 1999. Information technology-Coding of audio-visual objects-Part 2: Visual. International Standard ISO/IEC 14496-2:1999, ISO/IEC, Switzerland, December 16, 1999.

ITU and ISO/IEC., 1994. Information technology-Generic coding of moving pictures and associated audio information-part 2: Video. ITU-T Recommendation H.262 and ISO/IEC 13818-2 (MPEG 2 Video), March 25, 1994.

ITU and ISO/IEC., 2003. H.264: Advanced video coding for generic audiovisual services. ITU-T Recommendation H.264 and ISO/IEC 14496-10, Switzerland, May 30, 2003.

ITU., 1993. H.261: Video codec for audiovisual services at px64 kbit/s. ITU-T Recommendation H.261, International Telecommunication Union, Switzerland.

ITU., 1995. H.263: Video coding for low bit rate communication. ITU-T Recommendation H.263, International Telecommunication Union, Switzerland.

Li, H. and D. Doermann, 2002. Video indexing and retrieval based on recognized text. Proceedings of the IEEE Workshop on Multimedia Signal Processing, December 9-11, 2002, US Virgin Islands, USA., pp: 245-248.

Li, X.Y., X.H. Yang and Y. Yu, 2009. Research on individualized recommendation based on implicit feedback analyses. Comput. Eng. Design, 16: 3794-3796.
Direct Link

Lim, Y.K., S.H. Choi and S.W. Lee, 2000. Text extraction in MPEG compressed video for content-based indexing. Proceedings of the 15th International Conference on Pattern Recognition, Volume 4, September 3-7, 2000, Barcelona, pp: 409-412.

Lin, D.M. and J.B. Tsui, 2001. A software GPS receiver for weak signals. Proceedings of the IEEE MTT-S International Microwave Symposium Digest, Volume 3, May 20-24, 2001, Phoenix, AZ, USA., pp: 2139-2142.

Lin, K.H., Z.M. Hussain and R. Harris, 2004. Adaptive transmit eigenbeamforming with orthogonal space-time block coding in correlated space-time channels. Proceedings of the IEEE International Conference on Acoustics, Speech and Signal Processing, Volume 4, May 17-21, 2004, Australia, pp: 817-820.

Lu, W., 2007. Collaborative filtering algorithm and its application in personalized recommendation system. M.A. Thesis, Beijing University of Posts and Telecommunication, China.

Meng, X.F. and L. Chen, 2009. Collaborative filtering recommendation algorithm based on Bayesian theory. J. Comput. Applic., 29: 2733-2735.
Direct Link

Misra, P. and P. Enge, 2001. Global Positioning System: Signals, Measurements and Performance. Vol. 1, Ganga-Jamuna Press, India, ISBN: 9780970954404, Pages: 390.

Patil, S., S.K. Sangle, A.D. Yenare and R.S. Sonawane, 2015. Location based advertising android advertising application. Int. J. Recent Innov. Trends Comput. Commun., 3: 1357-1360.
CrossRefDirect Link

Tarokh, V., H. Jafarkhani and A.R. Calderbank, 1999. Space-time block codes from orthogonal designs. IEEE Trans. Inform. Theory, 45: 1456-1467.
CrossRefDirect Link

Taylor, R.E. and J.W. Sennott, 1984. Navigation system and method. U.S. Patent No. 4,445,118, April 24, 1984. http://ntrs.nasa.gov/archive/nasa/casi.ntrs.nasa.gov/19840014478.pdf.

Toubiana, V., A. Narayanan, D. Boneh, H. Nissenbaum and S. Barocas, 2010. Adnostic: Privacy preserving targeted advertising. Proceedings of the 17th Annual Network and Distributed System Security Symposium, September 11-18, 2009, India -.

Yang, J., M.H. Lee, M.Q. Jiang and J.Y. Park, 2002. Robust wireless image transmission based on turbo-coded OFDM. IEEE Trans. Consum. Electron., 48: 724-731.
CrossRefDirect Link