Measuring digital ranking on an online media platform

The eight-layered architecture system addresses the challenge of measuring digital ranking by integrating machine learning and AI to collect, process, and visualize data from multiple sources, offering accurate and scalable digital ranking measurements that enhance market reputation and reach.

WO2024228050A1PCT designated stage expired Publication Date: 2024-11-07SAYAL ANUJ +1
View PDF 2 Cites 0 Cited by

Patent Information

Application Number
PCT/IB2023/057575
Authority / Receiving Office
WO · WO
Patent Type
Applications
Current Assignee / Owner
Priority Date
2023-05-04
Filing Date
2023-07-26
Publication Date
2024-11-07

AI Technical Summary

Technical Problem

Current methods lack a comprehensive, cost-effective solution for measuring digital ranking on online media platforms, as they fail to integrate data from various sources accurately and efficiently, leading to incomplete and inaccurate evaluations.

Method used

An eight-layered architecture system that collects and processes data from multiple sources using machine learning and AI, featuring data source collection, preprocessing, feature extraction, network analysis, and visualization, with tools like Natural Language Processing and blockchain for data security and credibility assessment.

Benefits of technology

Provides accurate, scalable, and extensible digital ranking measurements, enhancing market reputation and reach by filtering out fake accounts and contents, and offering actionable insights through advanced sentiment analysis and visualization.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure IB2023057575_07112024_PF_FP_ABST
    Figure IB2023057575_07112024_PF_FP_ABST
Patent Text Reader

Abstract

An multi-layered system 100 and method to measure digital ranking on online media platforms include a data source collection unit 102, a data set unit 104 containing different types of data and parameters on which data source is extracted, a data retrieving unit 106 to retrieve data by setting a criteria by the user, a data privacy, security, and anonymization unit 108, a data storage unit 110 consisting of database servers for raw and processed data, a data pre-processing unit 112 for extracting features, a feature extraction unit 114 to extract data set by implementing algorithms, a network analysis, data mining and prediction unit 116 to provide results based on analysis algorithms, and a data visualization unit 118 containing parameters and filters on which user can evaluate data. System 100 is capable for data privacy, detecting and eliminating fake accounts and contributing to the false evaluation of digital ranking.
Need to check novelty before this filing date? Find Prior Art

Description

MEASURING DIGITAL RANKING ON A ONLINE MEDIA PLATFORMTECHNICAL FIELD

[0001] The present disclosure relates to the field of Online Platform including Online Media, Search Engine etc. More particularly, the present disclosure relates to measure digital ranking for an entity / product / service / personality on a online media platform for market reputation and reach in a gainful way.BACKGROUND

[0002] Background description includes information that may be useful in understanding the present disclosure. It is not an admission that any of the information provided herein is prior art or relevant to the presently claimed disclosure, or that any publication specifically or implicitly referenced is prior art.

[0003] In this era of information, the internet is widely used by the peoples for one or more reasons. Since peoples regularly discuss, communicate, and shares their data through the web they contribute lot of data, engagements on various types of platforms like Online media platforms that include Facebook, Twitter, Instagram, YouTube, and other Blogs, News platforms, web etc. for Facebooking, twittering, Blogging, posting videos on YouTube, providing feedback on newspaper articles or rating personality or services, are often seen making the internet an exceptionally basic method for correspondence in human life for both business and individuals.

[0004] Digital ranking is the website’s position on the different search engines, blogs, News, as well as on online media platforms. It is a brand’s presence on online media or the web. The importance of digital ranking is to make better web participation and to develop proper engagement between the business and the customers as the good digital ranking leads to improved web traffic. The digital ranking also contribute to increased circulation of ‘Social capital’ which we can be seen as a new form of currency and, more generally, value.

[0005] But there is no planform which provides extensive data analysis as a whole into one and provide digital ranking in specific domain. Since there are many data sets and many platforms for any brand / person / entity creating different types of datasets, filtering them, and then start analyzing for providing accuracy is very typical process.

[0006] There is, therefore, a need to overcome the above drawback, and limitations associated with defining digital ranking on platform especially on online media platform andprovide a simple, fast, cost-effective solution based on integrating Machine Learning algorithm with artificial intelligence.OBJECTS OF THE PRESENT DISCLOSURE

[0007] An object of the present disclosure is to provide a simple, fast, cost-effective solution for measuring digital ranking for an entity / product / service / personality on an Online platform.

[0008] Another object of the present disclosure is to provide an advanced technology-based solution using artificial intelligence and machine learning (ALML).

[0009] Another object of the present disclosure is to provide a solution to measure digital ranking that can be used by any entity / product / service / personality to meet their goal.

[0010] Another object of the present disclosure is to provide high performance, scalability, and high extensibility.

[0011] Another object of the present disclosure is to provide a solution for measuring digital ranking that help a personality and an entity to enhance their sentiment analysis market reputation and reach.SUMMARY

[0012] The present disclosure relates to the field of online media platform. More particularly, the present disclosure relates to measure digital ranking for an entity / product / service / personality on a online media platform for market reputation and reach in a gainful way.

[0013] According to an aspect, the disclosure is a system to measure digital ranking on an online media platform include a server including at least one database, and at least one processor communicatively coupled with memory instructions which, when executed, causes at least one processor to perform operation to collect a plurality of raw data from a plurality of data sources; distribute raw data to configure data set according to data type and parameters. Further, extract data type and parameters to automatically separate required data sets at an initial level using application programming interface and web crawlers; store separated data sets into database servers for meaningful extraction when required; perform pre-processing of data sets to achieve raw and processed data for extracting features; extract, by implementing algorithms, one or more features from the raw and processed data to prepare required data sets. The processor further perform to detect and eliminate fake accounts and contents by using the random forest algorithm with principal components analysis and datacorrelation; predict results using one or more analysis models configured to provide results based on the extracted data set through analysis algorithms; and display predicted results on a visualization unit comprising parameters and filters to evaluate the data for ranking.

[0014] In an aspect, the system is an eight-layered architecture comprising a data source collection unit, a data set unit, a data retrieving unit, data privacy, security, and anonymization unit, a data storage unit, a data pre-processing unit, a feature extraction unit, a network analysis, data mining and prediction unit, and a visualization unit.

[0015] In an aspect, the plurality of data sources is a web source comprising web news, blogs, and other news, and the other source is a onlinemedia source including social platforms including Facebook, twitter, Instagram, YouTube, and other platforms.

[0016] In an aspect, the pre-processing is done using tools comprising Natural Language Processing, Opinion mining, keyword-based functions, image emotion classifier, and others.

[0017] In an aspect, the API requests are sent with headers including information about the request.

[0018] In an aspect, the basis for feature extractions are sentiments, trends, the popularity score, demographic, mentions, influencers, and others.

[0019] In an aspect, the network analysis, data mining and prediction unit generates visualization features for digital impact, digital ranking, category ranking, trends, and sentimental analysis.

[0020] Another aspect of the disclosure is a method for measuring digital ranking on a onlinemedia platform include a server, the server include at least one database, and at least one processor communicatively coupled with memory instructions which, when executed, causes at least one processor to perform operation for collecting a plurality of raw data from a plurality of data sources; distributing raw data to configure data set according to data type and parameters; extracting data type and parameters for automatically separating required data sets at an initial level using application programming interface and web crawlers; storing separated data sets into database servers for meaningful extraction when required. The processer further perform pre-processing of data sets to achieve raw and processed data for extracting features; extracting, by implementing algorithms, one or more features from the raw and processed data to prepare required data sets; detecting and eliminating fake accounts and contents by using the random forest algorithm with principal components analysis and data correlation; predicting results using one or more analysis models configured to provide results based on the extracted data set through analysis algorithms; and displaying predictedresults on a visualization unit comprising parameters and filters to evaluate the data for ranking.

[0021] In an aspect, the method evaluating parameters also reaches for every post, every account for speed weighted reach on the plurality of data source platforms.

[0022] In an aspect, the detecting and eliminating fake accounts and contents contributing to the false evaluation of digital ranking.

[0023] Various objects, features, aspects, and advantages of the inventive subject matter will become more apparent from the following detailed description of preferred embodiments, along with the accompanying drawing figures in which like numerals represent like components.BRIEF DESCRIPTION OF DRAWINGS

[0024] The accompanying drawings are included to provide a further understanding of the present disclosure, is incorporated in, and constitutes a part of this specification. The drawings illustrate exemplary embodiments of the present disclosure, and together with the description, serve to explain the principles of the present disclosure.

[0025] In the figures, similar components, and / or features may have the same reference label. Further, various components of the same type may be distinguished by following the reference label with a second label that distinguishes among the similar components. If only the first reference label is used in the specification, the description is applicable to any one of the similar components having the same first reference label irrespective of the second reference label.

[0026] FIG. 1 illustrates an exemplary block diagram for the system with an eight layer architecture, in accordance with an embodiment of the present disclosure.

[0027] FIG. 2 illustrates a systematic flow diagram for a sentimental analysis used for feature extraction.

[0028] FIGs. 3A and 3B illustrates a conventional machine learning and deep learning models used for feature extraction in the sentimental analysis.

[0029] FIG. 4 illustrates a conventional sentimental analysis framework.

[0030] FIG. 5 illustrates an exceptional block diagram for proposed sentimental analysis, in accordance with embodiments of the present disclosure.

[0031] FIG. 6 illustrates an sentiment analysis block diagram for Twitter sentiment, in accordance with an embodiment of the present disclosure

[0032] FIG. 7 illustrates an exemplary method flow diagram, in accordance with an embodiment of the present disclosure.DETAILED DESCRIPTION

[0033] The following is a detailed description of embodiments of the disclosure depicted in the accompanying drawings. The embodiments are in such detail as to clearly communicate the disclosure. However, the amount of detail offered is not intended to limit the anticipated variations of embodiments. On the contrary, the intention is to cover all modifications, equivalents, and alternatives falling within the spirit, and scope of the present disclosure as defined by the appended claims.

[0034] In the following description, numerous specific details are set forth in order to provide a thorough understanding of the embodiments of the present invention. It will be apparent to one skilled in the art that embodiments of the present invention may be practiced without some of these specific details.

[0035] Embodiments of the disclosure relate to the field of online media platform. More particularly, the present disclosure relates to measure digital ranking for an entity / product / service / personality on a online media platform for market reputation and reach in a gainful way.

[0036] According to an embodiment, the present disclosure is an eight-layer architecture system and method to measure digital ranking on online media platforms include a data source collection unit containing details of a source, a data set unit containing different types of data and parameters on which data source is extracted, a data retrieving unit to retrieve data by setting a criteria by the user, data privacy, security, and anonymization unit, a data storage unit consisting of database servers for raw and processed data, a data preprocessing unit to process data for extracting features, a feature extraction unit to extract data set for features by implementing algorithms, a network analysis, data mining and prediction unit to provide results based on analysis algorithms, and a data visualization unit containing parameters and filters on which user can evaluate data. The system is capable of detecting and eliminating fake accounts and contents contributing to the false evaluation of digital ranking.

[0037] Referring to FIG. 1 where an exemplary block diagram for the system 100 is shown. The system 100 for measuring digital ranking on online media platform is an eightlayer architecture. The eight-layered architecture includes a data source collection unit 102, a data set unit 104, a data retrieving unit 106, a data privacy, security, and anonymization unit(108), a data storage unit 110, a data pre-processing unit 112, a feature extraction unit 114, a network analysis, data mining and prediction unit 116, and a visualization unit 118.

[0038] In an embodiment, the data source collection unit 102 is the first layer of the eight-layered architecture for measuring digital ranking on online media platform. The unit 102 include details of the data source from where the raw data has been collected. The sources may include Facebook, Twitter, Instagram, YouTube, Google news, Blogs, articles and news websites, and from other source. Whereas the data set unit 104, which the second layer include the details of the different types of data and parameters such that data types like Text, Image, Emoji’s, Polls / surveys, and parameters like Followers, Tweets and Retweets, influencers etc. on which extraction is done.

[0039] In an embodiment, the data retrieving unit 106 obtaining data from a social, web, or other considered sources. It is assumed that data is represented in a raw form, and there is ambiguity in data. Therefore, in order to retrieve the desired data, a set of criteria is made by any means by using crawlers, the data revival tools and their application programming interface (API). The API requests are sent with headers that include information about the request. When sending data with fetch (), it is required to specify the content-type, that tells the API for the data sent. All data that has been revived are a raw data at initial level.

[0040] In an embodiment, the function of data privacy, data security, and anonymization unit 108 ensure data privacy and security that focuses on anonymizing user data before it is processed and analyzed further. Here, the use of a plurality of data anonymization such as k-anonymity, 1-diversity, t-closeness are used to ensure differential privacy that all collected social data of users is de-identified and cannot be traced back to any individual user. The anonymization layer uses a combination of machine learning algorithms and deep neural networks to identify and remove any personally identifiable information (PIT) such as names, addresses, and phone numbers, as well as any other sensitive data that could be used to identify an individual.

[0041] Furthermore, the original identifiable physical user is replaced by a virtual user which retains all the important "thought and consumer profile" properties but can't be linked back to the real physical user. This approach to data privacy and security ensures that all collected social data is handled in a manner that complies with industry standards and regulations. By implementing this layer 108 of anonymization, the system 100 not only protects the privacy of users but also ensures that the insights derived from the analysis are accurate, unbiased, and ethical. Ultimately, the anonymization layer provides a new level ofdata security and privacy that has never been seen before in the field of onlinemedia analytics. The layer 108 demonstrates its ability to handle big data in a way that is both ethical and secure.

[0042] In an embodiment, the data storage unit 110 which is the fourth layer of eightlayered architecture, include database servers having Text and image database and a Statistical database, where all raw data and processed data are stored and on retrieval, it saves as a raw data on data servers. Servers have also been integrated with blockchain architecture to ensure transparency and immutability of the data collected. By leveraging the decentralized and tamper-proof nature of blockchain, the system 100 can create a secure and transparent ledger of all data transactions. Here all data collected from online media platforms is stored in an immutable and transparent manner, and provided a complete audit trail of data usage and access. Additionally, through blockchain authentication and authorization of users and data is secured, preventing unauthorized access and tampering.

[0043] In an embodiment, system 100 can provide a new level of trust and security for users, ensuring that their data is protected and handled in a transparent and ethical manner. This approach also aligns with emerging industry standards for data privacy and security, ensuring compliance and best practices are followed. Therefore, the integration of blockchain technology leveraging cutting-edge technologies to enhance data privacy, security, and transparency for the system 100 and staying ahead of industry trends, ensuring that the system 100 remains at the forefront of online media analytics through innovation.

[0044] In an embodiment, the data pre-processing unit 112 include the details of the different type of data and parameters on which data sources are extracted. In unit 110 data are processed exactly so we achieve the processed data which can be used for feature extraction. The tools used are a keyword-based processing functions, an image emotion classifier processing functions, an option mining functions, and a natural language processing functions.

[0045] In an embodiment, to address the issue of negative campaigns and fake news surrounding brands or keywords, an additional layer has been integrated into the architecture. This layer employs state-of-the-art machine learning algorithms to identify and analyze the sentiment and credibility of news articles and online media posts that mention the brand or keyword. System 100 further uses advanced Natural Language Processing (NLP) techniques, such as deep neural networks and transformer models like Bidirectional Encoder Representations from Transformation (BERT) and a third generation language model Generative Pre-trained Transformer-3 (GPT-3), to detect fake news and other forms ofmisinformation. It also utilizes machine learning models trained on large datasets of credible and non-credible sources, such as the logistic regression model and decision trees, to evaluate the trustworthiness of articles and posts.

[0046] In an embodiment, system 100 also uses a novel algorithm called the Credibility-Weighted Popularity Score (CWPS) to determine the true popularity of a brand or keyword. This algorithm takes into account the credibility of each source that mentions the brand or keyword, as well as the sentiment of the content. The Credibility-Weighted Popularity Score can be calculated using formula-CWPS = (P x C) / T where P is the total number of positive mentions, C is the credibility score of the source (ranging from 0 to 1), and T is the total number of mentions.

[0047] In an embodiment, by integrating this layer into the system 100 architecture, system 100 can accurately determine the true popularity of a brand or keyword, taking due consideration of any negative campaigns or false information enabling businesses to make data-driven decisions to protect and enhance their brand. Summarily, this advanced layer represents a ground breaking development in the field of onlinemedia analytics, demonstrating the ability of system 100 to handle complex, high-volume data with precision and accuracy, and to generate insights that are both technically sound and actionable.

[0048] In an embodiment, the feature extraction unit 114 extract the data sets for features that are to be extracted with implementing algorithms. To achieve this, proper data which sets has been used within specific algorithms for getting the features to show on the visualization unit 118.

[0049] In an embodiment, the a network analysis, data mining and prediction unit 116 is also combined with the network analysis and illustrate the systematic representation of the prepared data set which is created or stored after the processing of feature extraction unit 114. Based on this data the analysis and prediction unit 116 work and provide you the exact result. This unit 116 include all the analysis algorithms through which we are able to generate all the visualization features for digital impact (Online media, search, news, video etc.), Digital ranking (Brand value for all platform), category ranking (For all platform), Trends, and Sentimental analysis.

[0050] In an embodiment, the visualization unit 118 is the last and eight layer of the architecture. The unit 118 include the details of user visualization parameters, filters for evaluation of data.

[0051] FIG. 2 illustrates a systematic flow diagram 200 for a sentimental analysis used for feature extraction. The sentimental analysis is related with gathering data, investigating it, and decides the sentiment of the data which can be either negative or positive sentiment. It can be utilized in product purchase, before buying a product, the purchaser generally enquire about the opinion for the product from other purchasers. There are many techniques for the sentiment classifications as shown in FIG. 2. Here machine learning approach and lexicon based approach are used for sentiment analysis. The machine learning algorithms play an important role whereas supervised and unsupervised algorithms are utilized for the classification of comments as negative or positive.

[0052] In an embodiment, the lexicon-based approach is a semantic direction-based approach for slant investigation. It has been found that the lexicon-based techniques do not perform well on the characterization of sentiments, due to the opinion-based nature of the text, which requires all the more comprehension of the content.

[0053] In an embodiment, machine learning-based methods have been utilized besides lexicon-based approach and the feature extraction is utilized to distinguish a subset of features to accomplish different objectives for sentimental analysis.

[0054] FIGs. 3A and 3B illustrates a conventional machine learning and deep learning models used for feature extraction in the sentimental analysis. The machine learning uses specialized algorithms and deep learning networks for extracting features automatically from the text or images without the need of human intervention. This techniques is useful for accuracy improvement, improved data visualization, and increase in expalinability in model. The machine learning and deep learning are generally used on image data. Machine learning uses different models for getting required outputs whereas deep learning that works using different layers analyses the sentiments of a purchaser or visitor on online media giving an input in the form of rating, likes etc.

[0055] FIG. 4 illustrates a conventional sentimental analysis framework 400 using deep belief network with firefly-based sentiment analysis of a customer for a product review. The framework 400 based on paper “ A reinforced deep belief network with firefly-based sentiment analysis of customer product review “ by T. Sreenivasulu, an R. Jayakarthik, International journal of advanced Science and Technology, Vol. 29, No. 7(2020) pp 1504- 1516. According to the author, sentiment analysis is also referred as opinion mining which is review by the people about product purchased through online. The paper mentioned the process of sentiment analysis by performing mutual information-based feature selection andadopting deep belief network and functionality of the standard DBN is optimized by inferring the knowledge of firefly optimization to assign optimized weights on hidden nodes.

[0056] In an aspect, the paper used unstructured database consisting of vague information for data preprocessing where unwanted blank spaces are removed. This preprocessing converts data to numerical vectors where each vector corresponds to a review and entries of each vector represent the presence of a specified feature in that specific review. The modified mutual feature selection algorithm is proposed to remove the residual features from the selected features. The deep belief network (DBN) with multiple hidden layers having both directed and undirected edges. And the Restricted Boltzmann Machine (RBM) is an undirected propagative energy-based approach which includes of a single visible input layer, single hidden layer and a bias unit for interconnecting all layers. The firefly algorithm is with a flash light feature of the firefly population. According to the code of light intensity, the fitting role for the distance amid any two fireflies. The high brightness firefly is considered to have least distance and thus it is declared as the best solution for the objective function.

[0057] FIG. 5 illustrates an exceptional proposed block diagram 500 for feature extraction in sentimental analysis. In the proposed sentimental analysis is to measure digital ranking for online media platform defines that the Individuals share information, their experiences, and opinions with the world by utilizing OnlineMedia like online journals, newsgroups, discussion pages, forums, wikis, survey locales, interpersonal organizations, tweets, etc. The analysis is based on the fusion of different algorithms and mainly rely on the Deep Model. The major focus of the disclosure is the improvement of accuracy or performance of sentiment analysis. Contribution involves the optimal feature set selection and it is achieved through the feature’s reduction algorithm. Finally, handcrafted or conventional extracted features will be fused with the extracted deep features to get the strong features vector. Further, fused feature vector are fed to the multiple classifiers for the evaluation of the method performance.

[0058] In an embodiment, the trends analysis used average directional movement index (ADX) which is an indicator of trend strength in a series of prices of a financial instrument. The proposed disclosure uses indicator for technical analysts and is provided as a standard in collections of indicators offered by various trading platforms. The ADX for their internal coding and ADX is a combination of two other indicators developed by Wilder, the positive directional indicator (+DI) and negative directional indicator (-DI). The ADX combines them and smooths the result with a smoothed moving average.

[0059] In an embodiment, to calculate +DI and -DI, the price data consisting of high, low, and closing prices each period (typically each day) are used. One first calculates the directional movement (+DM and -DM):UpMove = today's high - yesterday's highDownMove = yesterday's low - today's low if UpMove > DownMove and UpMove > 0, then +DM = UpMove, else +DM = 0 if DownMove > UpMove and DownMove > 0, then -DM = DownMove, else -DM = 0 After selecting the number of periods (Wilder used 14 days originally), +DI and -DI are: +DI = 100 times the smoothed moving average of (+DM) divided by average true range -DI = 100 times the smoothed moving average of (-DM) divided by average true range The smoothed moving average is calculated over the number of periods selected, and the average true range is a smoothed average of the true ranges. Then:ADX = 100 times the smoothed moving average of the absolute value of (+DI - -DI) divided by (+DI + -DI)Variations of this calculation typically involve using different types of moving averages, such as an exponential moving average, a weighted moving average or an adaptive moving average. trader_adx(PECL trader >= 0.2.0) trader_adx — Average Directional Movement IndexDescription trader_adx ( array $high , array $low , array $close [, int $timePeriod ] ) : arrayParameters highHigh price, array of real values. lowLow price, array of real values. closeClosing price, array of real values timePeriodNumber of period. Valid range from 2 to 100000.Return ValuesReturns an array with calculated data or false on failure.

[0060] In an embodiment, the Interpretation can be taken as the disclosure ADX does not indicate trend direction or momentum, only trend strength. It is a lagging indicator; that is, a trend must have established itself before the check brand ADX will generate a signal that a trend is under way that ADX will range between 0 and 100. Generally, ADX readings below 20 indicate trend weakness, and readings above 40 indicate trend strength. An extremely strong trend is indicated by readings above 50. Alternative interpretations have also been proposed and accepted among technical analysts. For example it has been shown how ADX is a reliable coincident indicator of classical chart pattern development, whereby ADX readings below 20 occur just prior to pattern breakouts. The value of the ADX is proportional to the slope of the trend. The slope of the ADX line is proportional to the acceleration of the price movement (changing trend slope). If the trend is a constant slope then the ADX value tends to flatten out.

[0061] In an embodiment, Timing can be understood as there are various market timing methods have been devised using ADX. One of the best methods used in the disclosure is Alexander Elder Trading for a Living. One of the best buy signals is when ADX turns up when below both Directional Lines and +DI is above -DI. The customer would sell when ADX turns back down.

[0062] In an embodiment, the true range is defined for the first row is the difference between High and Low. That is 90 - 82, 8. Only truth shall set you free! While the previous statement is arbitrary, it is somewhat true too. If look at the high and low data, one might miss the price action happening between the consecutive days. This is where the true range comes into the picture.We take simply the maximum value of the following three components:(Current High - Current Low)Absolute value of (Current High - Previous Close) Absolute value of (Current Low - Previous Close) In that respect, we find the actual range of the price action between the days. For calculation, the True range of the above values, then for 2 December, the three values would be:(Current High - Current Low) = 95 - 85 = 10Absolute value of (Current High - Previous Close) = 95 - 87 = 8Absolute value of (Current Low-Previous Close) = (85-87)= Absolute value of (-2) = 2

[0063] Thus, since the true range is the maximum of the three values, it would be 10.

[0064] In an embodiment, the positive directional movement in the disclosure to help us gauge the uptrend of the market, it is pair with the Negative Directional indicator to derive real meaning from the indicators.

[0065] Furthermore. The Negative Directional Movement for checking the selling pressure, which could show up if the investors are ready to sell the stock at a price lower than the previous day. Thus, in the case of the Negative Directional indicator, the formula would be: If (Yesterday's Low - Today's Low) > (Today's high - Yesterday's High), then -DM = (Yesterday's Low - Today's Low) Otherwise, it's 0.

[0066] And the Smoothed values which is similar to moving average, smoothen out the fluctuations, if any, from the data. The period is defined as 5. This is arbitrarily chosen. The smoothed values of the positive directional movement. The first value is calculated by taking the sum of the first 5 values, i.e., (5 + 10 +15 + 20 + 25) = 75. The next value is calculated by taking the previous smoothed value and subtract the average from it. Finally, add the current value of the positive directional movement. In case of confusion, another example is given below.First smoothed positive directional movement value = 75.Average of the smoothed positive directional movement value = 75 / 5 = 15.Current positive directional movement = 30 (Recall that it is the sixth value)Thus, the second smoothed true range value = 75 - (75 / 5) + 30 = 75 - 15 + 30 = 90.

[0067] In an embodiment, the disclosure found out the Smoothed positive Directional movement as well as the Negative directional movement. On its own, either of the indicators has limited uses. But Wilder made use of both of them together so that their crossovers could be classified as a signal. Thus, for the positive Directional Indicator, we take the value of Smoothed Positive DM i.e., 75 and divide it by the true range, 70 and get the value 75 / 70 = 1.07 and as a percentage, it will be 107. Here, this is to note that the Smoothed negative DM was always 0, hence the Negative Directional Index Indicator will be 0

[0068] In an embodiment, the ADX Indicator for final calculations tells us the strength of the trend and not the direction. As mentioned above that we found out the positive directional indicator and the negative directional indicator. These would tell us the direction of the trend. But let us step back and try to see if we can get more information from the indicators. So, we have to check how far the two indicators from each other and this would give us a fair idea of how the market is behaving. If a situation where both the indicators are close to each other, then it means there could be a weak trend, if there is. But if there is alarge difference between the two, it means we have a strong trend. Wilder quantified this by giving the formula for the directional index (DX) indicator. It is as follows:

[0069] DX = (Difference of positive and negative directional index indicator) / (sum of positive and negative directional indicator)Thus, the DX value for 6 December would be DX = (75 - 0) / (75 + 0) = 1.As we are expressing them in percentages, it will be multiplied by 100 to give us the value: 1 * 100 = 100.And now, since we have to remove any fluctuations, we take the moving average of DX to get the Average directional Index (ADX) indicator.

[0070] In an embodiment, the Key Statistics Data in rapid internet presence or digital presence its paramount to measure one’s share of voice on digital space, there are tools which measures the mentions but no one is giving digital ranking and digital score and detailed measurable tools on below mentioned parameters Mentions Reach Engagement Search Trends Sentiment Analysis (Positive, Negative or Neutral) Followers etc. Impact

[0071] In an aspect, measuring top five hundred influencers on digital domain on Social Metrics Map for Digital Impact (Online media, Search, News, Video etc.), Digital Ranking (Include all platforms), Category Ranking (Include all platforms). Based on Different parameters we have developed an algorithm which provides best combination for the same. The algorithm is divided into two parts, 1. Sentimental Analysis (Positive, Negative or Neutral), 2. Statistical Analysis includes a. Mentions, b. Reach, c. Engagement, d. Search, e. Trends, f. Followers, g. Impact, h. Comments, News, Videos etc. and accordingly, points have been given and also score as compare to others to give them ranking.

[0072] In an embodiment, the Algorithm (Sentiment Analysis) (N) disclosed in the disclosure defines various steps as given below-Step 1— Initialize the TWStep 2— Evaluate the processStep 3 — while d < maxdata doIf(data-set=> positive datasetf])Store value in TWP DatasetIf(data-set=> negative datasetf])Store value in TWN DatasetEnd ifElseIf(data-set=> neutral datasetf])Store value in TWNt DatasetEnd ifEnd ifUpdate the TWn with new data setStep 4— end of whileStep 5— output the resultStep 6 - Take the value of (N)Abbreviation used in algo:-Tw — data set Twp Positive data collectionTwn Negative data collection TwNt Neutral data set7.. Based on the sorted output, 1st element will get the highest priority and the location for the same get from the location array.8. Print the output based on each category or as per the desired situation.9. Print the accurate combination of (T) from the matrix output from point No. 8Also, Brand Score Evaluation1. Begin2. Take value of (T) and value of (N)3. Evaluate 60% of (T) and 40% of (N)4. Brand Value (Bv) = 60% of (T) minus 40% of (N)5. Higher the value of N lesser will be the value of Brand.And, Technology / Platform Used1. Programming Language - PHP2. Framework - Laravel3. Database: MYSQL4. Front End: Bootstrap, Angular

[0073] FIG. 6 illustrates a sentiment analysis diagram 600 for Twitter sentiment. The twitter sentiment analysis 600, where some of the examples of tweets have been picked-up to show the positive and negative aspect using deep learning. The examples like “ I love weights & biases”, “Great visualisations on W&B- thank you Lukas & Team”, “I hate Al and W&B”, and “the product is exactly what I’ve been looking for” have been analysed for sentiments using deep learning where the tweets have been analysed using hidden layers. The result shows that the output has been segregated in positive tweets and negative tweets and identified “I hate Al and W&B” whereas all other tweets have been analysed as positive.

[0074] FIG. 7 where an exemplary method flow diagram 700 for measuring digital ranking on a online media platform include a server, the server include at least one database, and at least one processor to perform operation as per step 702 for collecting a plurality of raw data from a plurality of data sources. According to step 704, distributing raw data to configure data set according to data type and parameters. Further step 706 define extracting data type and parameters for automatically separating required data sets at an initial level using application programming interface and web crawlers.

[0075] In an embodiment, step 708 directs for storing separated data sets into database servers for meaningful extraction when required. The processer further, as per step 810, performs pre-processing of data sets to achieve raw and processed data for extracting features, and then extracting, by implementing algorithms, one or more features from the raw and processed data to prepare required data sets as per step 712.

[0076] Furthermore, step 714 define detecting and eliminating fake accounts and contents by using the random forest algorithm with principal components analysis and data correlation, then predicting results using one or more analysis models configured to provide results based on the extracted data set through analysis algorithms according to step 716. Finally, as per step 718 method 700 displaying predicted results on a visualization unit consisting of parameters and filters to evaluate the data for ranking.

[0077] In an embodiment, the method 700 evaluating parameters for every post, every account for speed weighted reach on the plurality of data source platforms and detecting and eliminating fake accounts and contents contributing to the false evaluation of digital ranking.

[0078] If the specification states a component or feature “may”, “can”, “could”, or “might” be included or have a characteristic, that particular component or feature is not required to be included or have the characteristic. As used in the description herein and throughout the claims that follow, the meaning of “a,” “an,” and “the” includes plural reference unless the context clearly dictates otherwise. Also, as used in the description herein, the meaning of “in” includes “in” and “on” unless the context clearly dictates otherwise.

[0079] Moreover, in interpreting the specification, all terms should be interpreted in the broadest possible manner consistent with the context. In particular, the terms “comprises” and “comprising” should be interpreted as referring to elements, components, or steps in a non-exclusive manner, indicating that the referenced elements, components, or steps may be present, or utilized, or combined with other elements, components, or steps that are not expressly referenced. Where the specification claims refer to at least one of somethingselected from the group consisting of A, B, C....and N, the text should be interpreted as requiring only one element from the group, not A plus N, or B plus N, etc.

[0080] While the foregoing describes various embodiments of the invention, other and further embodiments of the invention may be devised without departing from the basic scope thereof. The scope of the invention is determined by the claims that follow. The invention is not limited to the described embodiments, versions or examples, which are comprised to enable a person having ordinary skill in the art to make and use the invention when combined with information and knowledge available to the person having ordinary skill in the art.ADVANTAGES OF THE INVENTION

[0081] The present disclosure provides a simple, fast, cost-effective solution for measuring digital ranking for an entity / product / service / personality on a online media platform.

[0082] The present disclosure provides an advanced technology-based solution using artificial intelligence and machine learning (Al- ML).

[0083] The present disclosure provides a solution to measure digital ranking that can be used by any entity / product / service / personality to meet their goal.

[0084] The present disclosure provides high performance, scalability, and high extensibility.

[0085] The present disclosure provides a solution for measuring digital ranking that help a personality and an entity to enhance their sentiment analysis market reputation and reach.

Claims

We Claim:

1. A system (100) to measure digital ranking on a online media platform, the system (100) comprising: a server comprising at least one database, and at least one processor communicatively coupled with memory instructions which, when executed, causes at least one processor to perform operation to: collect a plurality of raw data from a plurality of data sources; distribute raw data to configure data set according to data type and parameters; extract data type and parameters to automatically separate required data sets at an initial level using application programming interface and web crawlers; store separated data sets into database servers for meaningful extraction when required; perform pre-processing of data sets to achieve raw and processed data for extracting features; extract, by implementing algorithms, one or more features from the raw and processed data to prepare required data sets; detect and eliminate fake accounts and contents by using the random forest algorithm with principal components analysis and data correlation; predict results using one or more analysis models configured to provide results based on the extracted data set through analysis algorithms; and display predicted results on a visualization unit (118) comprising parameters and filters to evaluate the data for ranking.

2. The system as claimed in claim 1 , wherein the system is an eight-layered architecture comprising a data source collection unit (102), a data set unit (104), a data retrieving unit (106), a data privacy, security, and anonymization unit (108), a data storage unit (110), a data pre-processing unit (112), a feature extraction unit (114), a network analysis, data mining and prediction unit (116), and a visualization unit (118).

3. The system as claimed in claim 1, wherein the plurality of data sources is a web source comprising web news, blogs, and other news, wherein the other source is a online media source comprising Social platforms including Facebook, twitter, instagram, YouTube, and other platforms.

4. The system as claimed in claim 1, wherein the pre-processing in the data preprocessing unit (110) is done using tools comprising Natural Language Processing, Opinion mining, keyword-based functions, image emotion classifier, and others.

5. The system as claimed in claim 1, wherein the API requests are sent with headers including information about the request.

6. The system as claimed in claim 1, wherein the basis for feature extractions in the feature extraction unit (114) are sentiments, trends, the popularity score, demographic, mentions, influencers, and others.

7. The system as claimed in claim 1, wherein the network analysis, data mining and prediction unit generates visualization features for digital impact, digital ranking, category ranking, trends, and sentimental analysis.

8. A method (700) for measuring digital ranking on a online media platform, the method (700) comprising: a server comprising at least one database, and at least one processor communicatively coupled with memory instructions which, when executed, causes at least one processor to perform operation for: collecting a plurality of raw data from a plurality of data sources; distributing raw data to configure data set according to data type and parameters; extracting data type and parameters for automatically separating required data sets at an initial level using application programming interface and web crawlers; storing separated data sets into database servers for meaningful extraction when required; perform pre-processing of data sets to achieve raw and processed data for extracting features; extracting, by implementing algorithms, one or more features from the raw and processed data to prepare required data sets; detecting and eliminating fake accounts and contents by using the random forest algorithm with principal components analysis and data correlation; predicting results using one or more analysis models configured to provide results based on the extracted data set through analysis algorithms; and displaying predicted results on a visualization unit (118) comprising parameters and filters to evaluate the data for ranking.

9. The method as claimed in claim 8, wherein the method (700) evaluating parameters also reaches for every post, every account for speed weighted reach on the plurality of data source platforms.

10. The method as claimed in claim 8, wherein detecting and eliminating fake accounts and contents contributing to the false evaluation of digital ranking.

Citation Information

Patent Citations

  • Mutually reinforcing ranking of social media accounts and contents

    US20170316519A1

  • Method and Apparatus for Collecting, Detecting and Visualizing Fake News

    US20210334908A1