In this analysis I want to:
Identify some cases of apps that do not require data and do collect it. These apps collect data even if it is not needed for their functioning.
The first step is to search for a functionality by using the topic modeling's score under a specific threshold, then, we have to consider the type of permissions
normally required by this type of app. finally, we consider these permissions that are not normally required and we identify group of desired apps.
For this analysis it could be good to generate a database with this finding. This table needs the following variables:
The name and id of the app, the app description, revenue model, popularity, and privacy policy, dates initial and last update
SDKS are also important!
Description of the app, why not.. just to understand what the app does.
Now we need to filter from those that have a second topic related to the permission due to the possible correlation of topics. This is done under two options, namely, stm effect relation, and data mining association rules.
First, STM computes topic modeling considering certain covariates which are going to be important variables to understand the topics meaning. These covariates are used to analyze certain differences in the text. For example, two classify politician's speeches, social scientists want to know the author's party, that could give a structure to this classification. Also, the analysis done considers the main parties, due to the convenience for understanding matters. The analysis considers variables that are not normally distributed and for this reason, it is not feasible the use of Pearson correlation. Being, Spearman rate the suggested method.
Second, one of the famous tools of Data Mining, classifies association of items under statistics describing the relationships. This computation is tedious to do without specific algorithms. Thus, the introduction of this method was after the insertion of the first algorithms into the computer. This tool let us find empirical statistics related to the data set that permits finding associated partners. What if one of the items is a probability? in this case, do we have to transform the prob to an event? or can we find an algorithm for this type?
The next steps include describing STM, and why computing this algorithm guided by covariates permissions, how studies used STM to find associations, what is the best statistic to measure correlation. What is the best method to perform association rules when the one is not an event but the probability of an event? Executing these steps contemplates reading papers, describing the methods in our draft, and computing and analyzing with the real data. The reading encompasses first, STM in social science papers, and business if some. These papers could even explain how to measure the association (i.e., if doing correlation is good), therefore, we do this reading before the rest. Second, topic modeling and other methods to association, again, this is found in the first selection of papers. So we prioratize topics found in the package website.
Do I need to have similar studies trying to find a link between collecting data and making money? making money with lies. What are their methods?
Identify some cases of apps that do not require data and do collect it. These apps collect data even if it is not needed for their functioning.
The first step is to search for a functionality by using the topic modeling's score under a specific threshold, then, we have to consider the type of permissions
normally required by this type of app. finally, we consider these permissions that are not normally required and we identify group of desired apps.
For this analysis it could be good to generate a database with this finding. This table needs the following variables:
The name and id of the app, the app description, revenue model, popularity, and privacy policy, dates initial and last update
SDKS are also important!
Description of the app, why not.. just to understand what the app does.
Now we need to filter from those that have a second topic related to the permission due to the possible correlation of topics. This is done under two options, namely, stm effect relation, and data mining association rules.
First, STM computes topic modeling considering certain covariates which are going to be important variables to understand the topics meaning. These covariates are used to analyze certain differences in the text. For example, two classify politician's speeches, social scientists want to know the author's party, that could give a structure to this classification. Also, the analysis done considers the main parties, due to the convenience for understanding matters. The analysis considers variables that are not normally distributed and for this reason, it is not feasible the use of Pearson correlation. Being, Spearman rate the suggested method.
Second, one of the famous tools of Data Mining, classifies association of items under statistics describing the relationships. This computation is tedious to do without specific algorithms. Thus, the introduction of this method was after the insertion of the first algorithms into the computer. This tool let us find empirical statistics related to the data set that permits finding associated partners. What if one of the items is a probability? in this case, do we have to transform the prob to an event? or can we find an algorithm for this type?
The next steps include describing STM, and why computing this algorithm guided by covariates permissions, how studies used STM to find associations, what is the best statistic to measure correlation. What is the best method to perform association rules when the one is not an event but the probability of an event? Executing these steps contemplates reading papers, describing the methods in our draft, and computing and analyzing with the real data. The reading encompasses first, STM in social science papers, and business if some. These papers could even explain how to measure the association (i.e., if doing correlation is good), therefore, we do this reading before the rest. Second, topic modeling and other methods to association, again, this is found in the first selection of papers. So we prioratize topics found in the package website.
Do I need to have similar studies trying to find a link between collecting data and making money? making money with lies. What are their methods?
Razaghpanah, A., Nithyanand, R., & Vallina-Rodriguez, N. (2018). Apps, trackers, privacy, and regulators: A global study of the mobile tracking ecosystem. Retrieved from http://eprints.networks.imdea.org/1744/
Cecere, G., Le Guel, F., & Lefrere, V. (2018). Economics of Free Mobile Applications: Personal Data. SSRN. https://doi.org/10.2139/ssrn.3136661
Comments
Post a Comment