When Lele was worrying about his mother's birthday present, he discussed it with his father. Dad mentioned that Mom likes to drink hand-cranked drinks, so Lele suggested to make bubble milk tea by herself. However, Dad reminded him that the key ingredients for making Jane Milk are not easy to find. Dad explained that the trial-and-error method takes a lot of time and the results are uncertain. Dad further proposed to use JMP for experimental design to find the bubble tea flavor that mom likes in a more efficient way.
Lele: Dad, Mom's birthday is coming, I want to give her a present, but I don't know what to give her so that she will like it.
Dad: Mom will be very happy with whatever you should give. Dad watched the delivery platform advertisement on TV and answered casually.
Lele looked at the advertisement and said with a flash of inspiration: Mom likes to drink hand-cranked drinks, or should I make him a cup of pearl milk tea?
Dad: Yes, but do you know how to do it? For example, the temperature of making black tea and how to cook pearls?
Lele: I can't, you help me.
Dad: There should be these production processes on the Internet, but there should be relatively little information on the temperature and cooking time, and everyone’s preferences are different. Since we want to give my mother a special birthday gift, we will make a special How about making her a cup of bubble tea exclusively for her as a birthday present?
Lele: Okay, but how do we find the flavor she likes? There should be an infinite variety of flavors in the world, so should we make an infinite number of cups?
Dad: That’s right, the method you said to try all the possibilities is the so-called exhaustive method or trial and error method, but this method will take a long time, and it may not necessarily be better. recipe combination. We should use JMP for experimental design.
what is JMP experimental design ?
Lele: What is JMP? What is experimental design? It sounds like an engineer doing an experiment. It feels so powerful!
Dad: JMP is a software that includes many statistical analysis functions, and you are right, one of its powerful functions can help us design efficient experiments. Moreover, JMP's experimental design can also set the type and range of factors according to your needs, add the possible effect model you think, and the number of experiments you can perform, etc., to help us construct the optimal experiment under the condition of ability.
Lele didn't seem to understand, and her eyes drifted to the unfinished Lego tree house.
Dad went on to say: All in all, we just have to tell him that we What are the options for black tea bags, what range can the water temperature be controlled, the desired time range for brewing tea, and we want it to do a few experiments, and it will help us design an experimental combination oh.
Lele: Oh, I see, like we hang out, Just like using driving navigation to help us design navigation routes, we can help us design the best route by telling him where we are going and where we are now. .
Dad: That's right, but navigation usually knows our goal. Is it to spend the least time? Or take the shortest path? Give me the most suggested combination according to the needs. And what we're going to do now is Data collection in the early stage, but using the set number of experiments, consider all possibilities as widely as possible, and collect its results. Then use this data to calculate the model and get the best prediction .
Lele: So now we are like driving navigation engineers, using JMP to design experiments for analysis, and then use the observed data to find out how to get to the destination the fastest?
Dad: That's right, that's it, let's use JMP to start our experimental design now, bubble tea engineer.
Lele and her father decided to use JMP's Design Of Experiment DOE (Design of Experiment) function to find out how to brew delicious black tea.
Dad: Let's find the recipe for making black tea first, and see what are the important influencing conditions.
Dad: Let's check the recommended methods of making black tea on the webpage, first find out which are the key conditions that affect the taste, and set the suggested conditions, so that we can narrow the scope of our experiment.
Therefore, Dad and Lele first found out the appropriate tea soaking temperature and soaking time, as well as the recommended water/tea volume ratio based on the information on the Internet. In addition, Dad found two brands of tea at a nearby store, and chose tap water or mineral water as the water for making tea. The key parameters are organized as follows (Table 1):
parameter
|
data type
|
value (range/type)
|
black tea brand
|
category
|
Brand A
|
Brand B
|
tea water
|
category
|
tap water
|
mineral water
|
Water/Tea Volume Ratio
|
continuous
|
20~60
|
Soaking temperature (⁰C)
|
continuous
|
70~100
|
Soaking time (m)
|
continuous
|
5~20
|
Table 1, key parameters and range of black tea
After searching for the key parameters and parameter ranges, Lele and Dad went back to the living room and sat down.
Lele: Then can we start making black tea now?
Dad: Not yet, because we still don't know the actual values of the parameters that need to be set in each experiment, and we still need a little helper to help us design the experiment.
Dad: We need JMP to help us design the combination of experiments.
So, Dad opened the JMP statistical software on the computer.
What is a custom experiment Custom Design
Dad: Look, if we need JMP to help us design experiments, you can click on the "DOE" column on the menu option. There are many DOE methods and related DOE functions that can be used here. If you are not so familiar with the principles and applicable objects behind each DOE, I suggest that you use custom design, that is, custom experiments, as a start.
Lele: Why use this custom experiment?
dad: The idea behind custom experiments is that optimization experiment , according to the conditions we gave him, including the key conditions we think are important, whether these key conditions affect each other, what kind of results we want to get, the number of experiments we can do, etc., to suggest our best Experimental mix.
Lele: Just like navigation.
Dad: This advantage is very suitable for ordinary people like us, who don't know DOE so well, but want to make a good experimental design.
Lele: I will become a DOE expert in the future. I am only a child now, so I won’t. I will when I grow up.
Dad: Okay, but even DOE experts often use customized experiments as the entry point for experimental design when the goal is to design the most effective experiments. So let's take a look now with a custom experimental design.
Use a custom experimental design, specifying parameter names, ranges, and number of experiments
Then Dad input the parameters and range into the custom design platform (Figure 1).
JMP Custom DesignJMP Custom Design
Figure 1, input response results and parameters to the custom design platform
The main effect of the parameter, the quadratic term of the continuous parameter and the effect term of the cross term of the continuous parameter were added to the model (Figure 2).
Figure 2, input the corresponding effect items into the model
Lele: What are quadratic terms and cross terms, and why should they be added?
Dad: The reason for adding the quadratic term effect is to confirm whether the key parameter will have a recurve phenomenon, that is, there is a value in the middle of the range that makes the result produce a maximum or minimum value. The addition of the cross term is to confirm whether these parameters will have a multiplicative effect on each other, that is, in addition to the influence of the main effect itself, certain conditions will have additional effects when combined.
Lele saw the picture below and asked, Lele: Is this the setting for how many experiments we want to do?
Figure 3, center point & repetition setting and experiment number input
Dad: Yes, here we can set whether to add repeated experiments or whether to add a center point to the experiment. The Number of Runs below can choose the number of experiments. Here, the minimum number of runs is 12 (Minimum). The JMP recommendation is 16 times (Default), you can also set an affordable number of experiments (User Specified), we still use the default 16 times (Figure 3).
Lele: Why 16 times? Isn't the more times the better?
Dad: Yes, but usually because of lack of time or money, we have a preset number of experiments that can be performed. And this is also the advantage of custom design. Custom design can design experiments according to this requirement, while other experimental designs must achieve the specified number of experiments. Another important reason is that you can’t drink tea, so I have to drink these 16 cups. Dad is already tense after drinking 16 cups, and he may not be able to sleep at night if he drinks a few more cups.
Generate the parameter combination for making black tea and complete the data collection
Lele and his father then completed the custom design and produced the parameter combination for making black tea (Figure 4).
JMP DOE參數組合JMP DOE parameter combinations
Figure 4, the experimental combination table
Then Dad and Lele completed the experiment according to this parameter combination table, and Dad scored in order and completed 16 scoring results (Figure 5).
Figure 5. Collection of corresponding experimental combination results
After drinking too many cups of tea, Dad hiccupped and said: We finally finished collecting data for the experiment.
use JMP Fit Model Perform regression analysis to make model predictions
Then, we can press "Model" in the script storage area on the upper left to do the subsequent analysis. JMP is very considerate to help us package the subsequent analysis in this program, that is, we can directly exist JMP Fit model on Platform to perform regression analysis , we can directly click "Model" on the DOE form to analyze (Figure 6).
Figure 6, the Model script on the DOE experiment combination data sheet
After pressing "Model", Dad went on to say: Here we can choose whether to add effect items to the model and choose the method of analysis, we will use "Stepwise" method please JMP Help us pick the right model (Figure 7).
JMP Fit ModelJMP Fit Model
Figure 7, the starting platform of the regression model
After pressing the mouse a few times, Dad completed the report (Figure 8), and pointed out the results of Effect summary and summary of fit on the computer with his fingers.
Dad: I exclude some items whose p-value exceeds 0.05 from the model in order to obtain a more accurate model, because statistically speaking, if the p-value exceeds 0.05, it means that the effect item has little influence on the model. Do you keep this? Effect terms do not differ. in addition, can be seen Summary of fit Here is the accuracy of the model (R-square) reached 0.997 , which means that this model has a strong ability to explain the sample points, which means that this is a good predictive model for these data .
JMP Fit ModelJMP Fit Model
Figure 8, regression model analysis report
Lele looked at it and said: Why is the P value of the soaking time greater than 0.05 but it has not been deleted? And there is an upward indicator next to it.
Dad: Because there are other sub-items that must be kept in the model, you can see that the soak time * soak time this item is less than 0.05 needs to be kept, as his parent item, the soak time also needs to be kept, and this upward indicator It's just telling you which ones are the parent projects that have this problem and must be kept.
Dad continued to scroll the mouse wheel, presenting the bottom half of the report.
inspect Prediction Profiler Reports observed changes for each parameter on the results
Dad went on to say: We went down to find the report. Prediction Profiler( marker) Report , here the model is presented with interactive pictures, which can help us understand faster How does changing each parameter change the flavor? . You can see the name of the parameter below, its axis (X axis) represents the parameter value we set, and the left axis (Y axis) represents the predicted value of the taste score. If you want to know the scores of different conditions, you only need to adjust the parameter value below to the value you set, and you can get the predicted taste result. For example, I choose A tea, soak it with tap water, the amount of water/the amount of tea I choose 40 for the ratio, 80 for the soaking temperature, and 15 for the soaking time, and when I change it like this...you see, forecast result It came out, only 4.83 points (Figure 9).
JMP Prediction ProfilerJMP Prediction Profiler
Figure 9, Prediction Profiler
Dad said while dragging the mouse to change the parameter settings, Lele's eyes widened, as if feeling incredible.
Lele: He will directly change the predicted taste score on the picture according to your adjustments, so cool.
Dad: Yes, this interactive function is easier to operate and makes it easier for everyone to understand the model. Let's take a closer look at this graph. In fact, this graph can roughly tell you which direction to adjust to get a higher taste score, which means it tastes better. For example, let's look at the first tea map, B Tea is better than A Tea is high, can know B Tea has a ratio A Trend of delicious tea , similarly, Tea made from mineral water tastes better than that made from tap water ,as well as The soaking time is about 10~15 It will be better in minutes , for you to play around with.
JMP Prediction ProfilerJMP Prediction Profiler
Figure 9, Prediction Profiler
Lele couldn't wait to start pulling various parameter settings. After playing for a while, he raised a question.
Lele: Dad, look, if I change the ratio of water volume/tea volume, other parameters will change accordingly, some will become higher and lower as a whole, and some will change their shape slightly, which is so strange.
Dad: Hey, you are amazing, you have noticed this small change. These small changes come from The effect of adjusting parameters , so other parameters will increase in shape. If you find that with the change of the original parameter value, the shape of other parameters has obvious variation. items have a strong impact. In addition, you can see that the trend line of some parameters is not a straight line, but a curved line. If there is a maximum or minimum value, it means that the effect of its secondary term is very influential.
Lele nodded half-understood, and then asked: So I can predict all possible outcomes in this way?
Dad replied: This is a good question. What we just did was to ask JMP to provide us with an estimated model based on the data we collected, but these data are based on broad and uniform points within the range we defined at the beginning. Try to look at the performance of all points within the range as much as possible, but for parameter combinations outside the range, the predictive ability of this model will be poor, so it is better for us not to make predictions outside the original experimental design range as much as possible, but for parameters within the range For most of all parameter combinations, we can predict an accurate result with this model.
Next, we are going to do today's key steps: Find the best combination of parameters and conditions , you can adjust the parameter settings on the scriber to see what kind of combination you find the highest score.
Lele nodded, and there was another burst of parameter setting adjustments on the Prediction Profiler.
exist Prediction Profiler , to find the best combination of parameters
Lele: Dad, I found one with a score of 9.47. I chose B tea leaves and soaked them in mineral water. I chose 60 for the ratio of water volume to tea volume, the soaking temperature was 89.5, and the soaking time was 11.9. The result is a taste score of 9.47, which is very good. Great.
Dad: Not bad, already very high, do you think there will be higher scores?
Lele: Maybe there will be, because I haven't set all of them to see his scores.
Dad: Now is the time to show the magic of JMP. You see, I can enable set desirability on the JMP Prediction Profiler and set goals. choose Maximum To set the result as hopeful, that is, the bigger the better . and use Maximize Desirability function to obtain the best combination of parameters for the completed goal (Figure 10). Or select Maximize and Remember to find the best combination of parameters and record it in the report.
Figure 10, (left, middle) Set Desirability function to set the willingness value, (right) Maximize Desirability
Dad presses Maximize and Remember on the Profiler to quickly find the best combination of parameters and record it on the report (Figure 11).
Figure 11, Maximize and Remember results
Lele: This is too fast and convenient!
Dad: Yes, Usually data analysis takes the most time to collect data. JMP It can save more analysis time, so many companies that need to use data for statistical analysis will use this software to speed up the analysis time . In fact, comparing the parameter combination you manually obtained is only a little bit worse than the best combination, you are also very good. In addition, continuous parameters, such as soaking temperature, 91.02894, are actually not easy to control. Maybe we can use 91 or 90 to replace the setting.
Lele and Dad then discussed the setting of various parameter conditions, and finally used B tea leaves, mineral water, the ratio of water volume to tea volume was 55, the soaking temperature was 90°C, and the soaking time was 12.5, which was the final key parameter condition setting (Fig. twelve).
Figure 12. The best prediction result according to the experimental design
Lele: Yeah! Then shall we make milk tea and pearls next?
Dad looked at the time, and then said: "Mom is coming back soon. If we want to treat it as a surprise for my mother's birthday, let's put away the experimental equipment as soon as possible, or we will be discovered." Next time we will experiment with milk tea and pearls to find a good formula combination, which will use other DOE techniques.
Lele said happily: "Okay, then let's collect it as soon as possible, I am looking forward to it."
Dad and Lele rushed to clean up the experiment before mother came home, but a small DOE seed slowly germinated in Lele's heart.
>>The next episode will share how to make delicious pearls through the JMP DOE experimental design platform. <<
Get more DOE design of experiments related resources:
