Wardz….

In each of these cases, a random selection from the database would have provided more efficacious results than applying the association rule. With association rules, one needs to keep in mind the prior probabilities involved. To illustrate, we now ask Clementine to provide us with a priori association rules, but this time using the confidence difference as the evaluative measure. Here, rules are favored that provide the greatest increase in confidence from the prior to the posterior. The results are shown in Figure 10.7. (more…)

Whenever you have problems to buy a house, a home loan service will definitely give you a lot of way outs in the end. But whether you recognize it or not, there will be a time when you fell like going down and become unable to finish the loan very well right away. If you have been trapped inside this kind of problem, there is no need to worry about it by the way. How could this even be possible? (more…)

?Association rules need to be applied with care, since their results are sometimes deceptive. Lets look at an example. Turning back to the a priori algorithm, we asked Clementine to mine association rules from the adult database using 10% minimum support, 60% minimum confidence, and a maximum of two antecedents. The results are shown in Figure 10.6. Consider, for example, the third association rule from the bottom, If Work Class = Government, then sex = Male, with 62.7% confidence. Marketing analysts interested in (more…)

?As mentioned above, GRI can handle numerical inputs as well as categori- cal inputs. We illustrate this using Clementine on the adult data set, instructing the GRI algorithm to accept both numerical variables and categorical variables as possi- ble antecedents (although, still, only categorical variables are possible consequents). The results, for minimum support and confidence criteria similar to those above, are
shown in Figure 10.5. (more…)

For rules with more than one antecedent, p(x) is considered to be the probability of the conjunction of the variable values in the antecedent. As usual, the user specifies desired minimum support and confidence criteria. (more…)

The structure of association rules, where the antecedent and consequent are both Boolean statements, makes them particularly well suited for handling categorical data, as we have seen. However, what happens when we try to extend our association rule mining to a broader range of data, specifically, numerical attributes? Of course, it is always possible to discretize the numerical attributes, for example, by arbitrarily defining income under $30,000 as low, income over $70,000 as high, and other income as medium. Also, we have seen how both C4.5 and CART handle numerical attributes by discretizing the numerical variables at favorable locations. (more…)

Sales are important part of a company. If you have best sales team and marketing you will have best result that will benefits the whole company or business. That’s why you need to do many things to improve the ability of the whole sales people in your company or business. If you need a help to handle it, you can use tools to help you. You can use crm. (more…)

Thus far, we have examined association rules using flag data types only. That is, all of the vegetable stand attributes took the form of Boolean 0/1 flags, resulting in the tabular data format found in Table 10.3, reflecting a straightforward market basket analysis problem. However, association rules are not restricted to flag data types. In particular, the a priori algorithm can be applied to categorical data in general. Lets
look at an example. (more…)

Compare Table 10.7 with Figure 10.1, the association rules reported by Clementines version of the a priori algorithm, with minimum 80% confidence, and sorted by support confidence. The first column indicates the number of instances the antecedent occurs in the transactions. The second column, which Clementine calls support, is actually not what we defined support to be in this chapter (following Han and Kamber[1], (more…)

Unique products and complete services are not enough in order to make a company gain good sales and earn profit. Company also needs good marketing strategies and plans to get into customers mind especially new company with new product. The company must able to communicate the idea why the product is produced. If the idea of the product can be communicated fluently in interesting way, the company will successfully get customers attention. When the customers begin to know that there is new product, they will curious with its features. In order to introduce a product, the company must choose suitable way which is able to deliver the message exactly to the customers. (more…)