Practice of Teaching Creative Writing with Text in a High-quality Context for College Students Based on Predictive Analysis Ability of Big Data

INTRODUCTION: With the advent of the digital age, big data has become essential to all walks of life. In digital education, it has become an indispensable part, which can cover the students' learning progress, learning hours, practice scores, course information, classroom interaction information and so on, which makes the teaching process fully covered and a one-stop learning experience. Big data is essential to promote China's education to the scientific, systematic, informatization and customization continue to move forward. OBJECTIVES: To further improve the writing ability of college students in China, the systematic learning of big data is integrated into the daily teaching and life of the university so that college students can better embrace this significant data era. METHODS: This paper discusses the analysis of the current situation of college students' accompanying writing, the problems and improvement methods of college students' accompanying creative writing, the systematic elaboration of big data, and the method of predictive analysis of big data to explore the practice of teaching college students' accompanying creative writing in a high-quality context based on the ability of predictive analysis of big data. RESULTS: With the advent of the digital age, big data has become an essential part of various fields. As an even more indispensable part of the digital education field, it can cover students' learning progress, learning time, practice results, course information, classroom interaction information, Etc., indeed covering the entire education program for a one-stop learning experience. CONCLUSION: For language teachers in colleges and universities, improving the ability of predictive analytics under big data and then integrating it into the teaching practice to cultivate college students with high-quality creative writing skills based on the ability of predictive analytics of big data is the current priority, and is also the goal of colleges and universities to cultivate talents.


Introduction
With the continuous modernization, customization and personalization of the university education system, the requirements of significant universities for creative writing accompanying texts for college students have been increasing [1]. Coupled with the knowledge explosion in the age of information technology, the threshold of acquiring knowledge in various industries has been continuously reduced, but the processing and analysis ability of massive knowledge and information has yet to be correspondingly improved, which is the biggest obstacle to high-quality writing and Xu • Li 2 accompanying creative writing of college In the era of facing massive data, big data technology comes into being and gradually spreads. Big data has incomparable advantages over inherited information processing methods when dealing with massive data -fast computing, fast information extraction, accurate feature marking, Etc. The college students of the new era urgently need to understand extensive data analysis methods, master data processing operations and use big data for predictive analysis, Etc. However, the restructuring of the college education system has an overall lag, which requires college students to learn the relevant content independently by consulting materials. The so-called big data is a massive amount of data, which also includes the methods of processing the vast amount of data and the methods of analyzing the vast amount of data. Traditional writing is often limited to limited information (pictures, words, cartoons, news facts, Etc. While writing under big data covers all dimensions of social production and life and is all-inclusive. Moreover, the description of affairs is based on the support of massive data, which is far more objective and persuasive than traditional writing and closer to reality. I believe that college students in the information age must master the ability to use and analyze big data, and then combine it with their personal literary quality and writing skills, and apply it to the accompanying creative writing so that they can write good articles, express their inner thoughts and laws through words, and improve their literary quality. More importantly, through many exercises, students can master a way of thinking to analyze and understand affairs. Through this way of thinking to see the world, they will be more objective and rational. They will cultivate college students with excellent quality, divergent thinking and creativity to build and develop the country. It is also clearly mentioned in the Modernization of Chinese Education 2035 that it is necessary to cultivate college students who have an open vision and master the understanding and analysis of things, to use various modern and informational resources flexibly, to enrich writing creativity, to cultivate students' enthusiasm for writing, and to cultivate a way of thinking in writing [2].

Research Background
Big data is one of the most critical resources of society, i.e., the collection of massive data. Depending on the context of the time, "big data" is not the same [3]. For example, from the "Shen Nong Bai Cao Jing" (a total of 365 kinds of medicinal herbs), which was compiled by medical doctors in the ancient and pre-Qin dynasties, to the "Ben Cao Gangmu," which was compiled by Li Shizhen in the Ming dynasty after 27 years, with 11,096 kinds of prescriptions and 1,892 kinds of medicinal herbs. Each era is different, and the corresponding big data is also different. Today, riding on the express train of the information age, the amount of data generated has increased exponentially yearly, producing the first truly massive amount of data in human history (the amount of data in 2021 alone has already exceeded the sum of the twentieth century). Generating Big Data is also accompanied by feature extraction, processing and analysis of the data. Another important feature of big data has also arisen in data analysis: a predictive analysis based on big data, which has become a necessary discipline. Through the analysis of massive data, a predictive assessment of individual or group behavior is made, and based on the massive amount of data, the accuracy of this assessment is very close to the actual probability. This process generates a variety of data, and the analysis generates new data, which, when added together, increases year by year, as shown in Figure 1. The modern university teaching system increasingly relies on informatization in the new era. In the background of informatization, the difficulty of acquiring knowledge is reduced. The threshold of learning various kinds of knowledge is reduced to the lowest, an era of "knowledge explosion" [4]. In this era, new requirements are put forward for informatization teaching in colleges and universities, not only to give full play to the subjective initiative of teachers but also to give full play to the data knowledge existing in the network in the informatization era and to use the big data in the informatization era as the pillar to give full play to the characteristics of "teaching according to the material" and to make corresponding personalized and customized teaching plans for each student. The teaching system that seemed impossible in the past has become a reality. Based on the application of information technology platforms, processing and analysis of big data, we can easily arrange the teaching plan. In order to achieve the best teaching effectteaching each other also. Also, through the information technology platform, students can view messages, notifications and assignments in real-time, and even synchronize to the teacher's correction of assignments and provide real-time feedback on the changes after the teacher has completed the correction, all of which are based on the information technology teaching system and the support of the country's various information technology infrastructures. Among them, various data resources play an essential role as intermediaries, as shown in Figure 2. The mediating role of data in informatization teaching Students acquire various data knowledge from the network and teaching platform. At the same time, they return various allowed collected data to cyberspace and teaching platforms, and they are also the content creators of knowledge while learning knowledge. Teachers send various data knowledge to students through the teaching platform on the one hand and get students' feedback data through the teaching platform on the other hand. Through the analysis and processing of students' feedback data, teachers can adjust their teaching arrangements and teaching plans in time and finally get involved in students' teaching planning as a whole, and even compile and organize the data to determine the direction of the extensive research study of the course. For the information technology teaching platform, it is then faced with a massive amount of data at the network end, the student end and the teacher end, and it is also necessary to use the method of extensive data analysis and processing to properly save and use the data for optimizing the learning experience and iterating and updating the platform functions, and even adjusting the development direction of its platform. Through extensive data analysis, the direction of subject development can be predicted and fed back to the teacher's side, forming a virtuous circle. For the web end, the amount of data interactions generated is enormous and, in contrast to the former, is massive data. For college students writing, writing is the most essential quality requirement. Since human beings have written, writing has become an important carrier to observe, describe, and perceive the world, a universal requirement for college students under the requirements of the new era [5]. College students who are proficient in writing use writing to record their lives on the one hand and express their perceptions and feelings on the other. This can not only improve the quality of humanities but also teach a way to discover the beautiful details of life, which can effectively release the pressure of the usual life and record the state of mood at this moment with the method of writing. Facing fast-paced study and life, it is an excellent way to regulate. Moreover, writing can cultivate students' imagination and creativity; these are the new social development background, the new comprehensive quality requirements for students, for language teachers in colleges and universities, improve the ability of predictive analysis under big data, and then in the teaching practice to integrate into the training of students with the ability of predictive analysis based on big data, high-quality creative writing ability of college students is the current It is also the goal of universities to cultivate talents. Teachers should also put forward improvement opinions based on predictive analysis of big data according to the problems exposed by students in writing practice so that they can be targeted in the face of various problems and promote the steady progress of college writing teaching, forming a positive feedback adjustment mechanism of teaching-feedback-improvement. The positive feedback adjustment mechanism of teaching-feedbackimprovement is shown in Figure 3.

Figure 3 -Diagram of teaching feedback mechanism
This feedback mechanism objectively improves the effect of writing teaching to a certain extent and provides a paradigm for college teachers in writing teaching. First, the teacher needs to be keenly aware of current social events and, through his or her initial collection of information, publish independent writing material on which the students write creatively. After receiving the independent writing material from the teacher, students look for writing angles through their understanding of the material; each of them chooses an angle they like or are interested in and then divide the work into groups to carry out material data collection, a step that requires students' ability to collect material data on their own, i.e., the first step in developing the ability to collect information on big data [6]. Then, students share the collected materials and data and analyze the data according to their perspectives, which is the step that requires students' predictive analysis ability of big data and must have a solid ability to analyze and process data in order to grasp the rhythm of the lines and the conclusion accurately. This step of group work writing is not only a requirement for college students' considerable data collection ability and extensive data analysis processing ability (predictive analysis ability based on big data) under high-quality writing, but also the process can cultivate students' communication, division of labor and cooperation and sharing qualities, which is the most critical step in writing teaching and the core step. The next step is for the teacher to focus on revising the students' creative writing. In this process, the teacher must control the general direction of the creative writing Xu • Li 4 content based on the teacher's rich writing experience. This control is based on the teacher's rich writing experience. The teacher will summarize the common and representative problems in students' writing. Finally, the teacher distributes the results of the corrections and then brings the students together for a question-and-answer session, in which the general problems are answered. The problems raised on the spot are analyzed in depth. The feedback mechanism is divided into four parts as follows.
(1) Feedback on the level of understanding of the material After the teacher releases the independent writing materials, students read and understand the materials, look for their writing directions, then analyze their writing perspectives, including the position, viewpoints, objects of analysis and conclusions obtained from their perspectives, Etc., and then give feedback on the writing directions together with the analysis report to the teacher, who assesses the writing directions, and if the assessment result is passed, they can prepare for the subsequent step division of labor. If the assessment result is not passed, the students revise the directions according to the revisions and then continue the feedback until the directions are passed [7].
(2) Feedback on the data collected by the group division of labor The method is to sort all the objectives in order of importance and then organize and categorize the data according to the order of data collection, in which there are strict requirements for data collection, analysis, and processing. The group members share the data and get preliminary conclusions by processing and analyzing the data and then giving feedback on the results to the teacher. The teacher assesses the students' conclusions according to their experience and sensitivity to the data. If the assessment result is passed, the student is ready to proceed to the next step. If the assessment result is not passed, the student modifies the result according to the modifications and then continues the feedback until the conclusion is passed.
(3) Feedback on the analysis of the team teacher in the centralized correction to identify common problems This method generates feedback from several students, which is collected and organized by the class committee and given to the teacher, including additions to what the teacher perceives as prevalent problems, check-ups for problems not identified by the teacher, and priority questions for specific problems. When given to the teacher, this feedback was a tremendous help for the subsequent focused Q&A.
(3) Feedback on the content of the centralized questionand-answer session The last step is to give feedback on the content of the centralized Q&A, and this time it is the last opportunity for the teacher to analyze, answer, and solve the questions as thoroughly as possible about the content of the accompanying creative writing [8]. Compared to the previous Q&A, it can be seen that the effectiveness and satisfaction of this Q&A were significantly higher due to all the modifications and feedback mechanisms prior to the Q&A, as shown in Figure 4.

Materials and Methods
Review of the theoretical basis of big data predictive analytics: With the advent of the digital era, big data has become an essential part of all walks of life. In digital education, it has become an indispensable component that can cover students' learning progress, learning hours, practice results, course information, classroom interaction information and so on, genuinely achieving full coverage of the teaching process and a one-stop learning experience. Big data is one of the most essential resources in society, i.e., a collection of massive data. The study of big data and the processing and analysis of these massive data are the developmental qualities that every university student is required to understand and master in the era of development, and this developmental quality has become a crucial part of writing teaching. Based on images: Big data, due to its vast volume, cannot intuitively see the meaning represented behind the data, so the research about big data is based on images to illustrate the problem, and only intuitive images can lower the threshold to pass up the information, in daily production life, everyone cannot understand the knowledge related to big data, through intuitive images, let the information circulate with low threshold, which is an effective way to reduce social costs and improve the efficiency of communication, as well as in teaching writing [9]. That is why extensive data analysis and fundamental principles are based on images. Algorithms are methods of analyzing and processing data. They are the core and critical steps of the predictive analysis of big data. The selection of different algorithms produces different results, so understanding the principles and suitable scenarios of various algorithms is an essential part of teaching writing. For example, these algorithms: principal component analysis, typical correlation analysis, stepwise regression, hierarchical cardinality analysis, and frequency statistics all have The flexible choice of different algorithms is considered to be a test of teachers' writing teaching practice and college students' quality accompanying writing based on predictive analysis of big data. Prediction and analysis based on big data. It is the most essential part of ample data use and the core value of big data. According to different analysis methods, the predictions are different, and we need to choose the correct algorithm to perform predictive analysis. After establishing a data analysis model, predictive analysis can be performed on the newly generated data, and the more significant the amount of data it has, the higher the reliability and accuracy of the model [10]. This paper uses the decision tree approach to process big data and build a decision tree model to perform predictive analysis of unknown data. Algorithm description: decision tree algorithm model is a class of methods used to solve regression and classification problems, and standard algorithms include ID3, CART, C4.5, Etc. We can generate a decision model with an excellent fit and good generalization ability through decision tree learning. The decision tree algorithm has significant advantages over other algorithms dealing with unknown data. The usual model of a decision tree model (let us take a dichotomous classification task as an example) is shown in Figure 5. In a decision tree for machine learning using big data, we want the data in the branch sub-nodes of its decision to be as much as possible a type of data that can be abstracted to make that branch word node purer and purer. Then the decision tree can be efficient for decision classification and gain predictive power from considerable data training. Thus, prediction can be performed efficiently. Our evaluation index for this prediction effect is "information gain," and under "information gain," let us first understand "information entropy" [11]. Information entropy is one of the most widely used metrics to measure the "purity" of our data set. If we assume that the proportion of the kth class of data set in the current data set D is pk (k=1, 2, 3, ...), then we call it the "information entropy" of that class of data set. "information entropy." (1) The smaller this value is, the greater we consider the "purity" of the data set of that category. We generally perform the binary classification task for our ordinary decision tree, and we have y=2. However, if we already know that a value measures the uncertainty of sending events, then we call this value "information content." For informativeness, it has the following characteristics.
(1) The value cannot be negative.
(2) The uncertainty function f(x) for the data is a monotonically decreasing function of our probability function p.
(3) Summability, i.e., the value of all the informative quantities of two independent types add up to the informative quantity of the two types grouped into the same category, as shown in equation (2).
(2) For a function, f(x) that satisfies all three of the above at the same time is a negative logarithmic function, i.e. (4) Assume that the discrete attribute a has v possible values {a1,a2..., av}. If we use the information entropy of the discrete attribute a, the information entropy of the discrete attribute a can be calculated. , av}, if a is used to partition the sample set D, v branch nodes are generated, where the vth branch node contains all the samples in D that take the value av on the attribute a., denoted as Dv. We can calculate the information entropy of Dv according to equation (4), and then consider the different number of samples contained in different branch nodes, and give the branch node a weight |Dv|/|D), i.e., the number of samples The more the number of samples, the greater the influence of the branch node, so we can calculate the "information gain" obtained by dividing the sample set D with attribute a.

( )
In general, we take out a significant information gain value, meaning we get a significant "purity improvement" by classifying the data category. When we train the model Xu • Li 6 based on big data, the more significant the amount of data and the wider the distribution of data samples, the stronger the prediction ability of our decision tree based on big data. In this paper, we investigate the practice of teaching creative writing in a high-quality context for college students based on the predictive analysis ability of big data.

Results and Discussion
Mechanistic analysis: This section compares the influence of big data's predictive analytic ability on teaching creative writing in high-quality contexts among college students based on theoretical knowledge and concrete practice. This paper analyzes the correlation between college students' accompanying creative writing and teaching practices and the influence of characteristic and context-specific factors on the frequency of college students' high-quality accompanying creative writing behavioral practices on multiple types of teachers' writing instruction. This paper provides an in-depth analysis of how big data can be used for predictive analysis and how to develop predictive analysis skills of college students based on big data through a multiple decision tree (information gain) algorithm [12]. With the advent of the digital age, big data has become an essential part of all walks of life. In digital education, it has become an indispensable component that can cover students' learning progress, learning hours, practice scores, course information, classroom interaction information, and so on, genuinely achieving full coverage of the teaching process and a one-stop learning experience. Big data has been instrumental in pushing China's education forward toward scientific, systematization, informatization and customization [13]. In recent years, big data has been increasingly applied in teaching and learning, and predictive analysis based on big data has become an increasingly important ability for contemporary college students. It is used throughout all stages of learning. With the development of the times, the time and space scope of college students' essay writing is broad, and in such a context, the model of college students essay writing also requires changes to embrace this era of big data better. Big data predictive analysis has become an essential part of essay writing, and only by mastering and using big data proficiently to conduct analysis can we write forward-looking essays. In order to further improve the writing skills of our university students, systematic learning of big data should be integrated into the daily teaching and life of the university. In this paper, through the analysis of the current situation of college students' essay writing, the problems and improvement methods of college students' essay creative writing, the systematic elaboration of big data, and the predictive analysis method of big data, we explore the practice of teaching college students' essay creative writing in the context of high quality based on the predictive analysis ability of big data.
Teaching creative writing in the context of high quality based on predictive analysis of big data can be an excellent way to cultivate the thinking of college students. The structure of creative writing, which is ready to see and write and task-driven, also effectively avoids the dilemma that students have nothing to say. In addition, different writing angles can be derived from different angles of analyzing the material, effectively improving students' divergent thinking and preventing them from having the same ideas. The summary at the end of the course is a series of words to form an essay. Teaching creative writing in the context of high-quality college students based on predictive analysis of big data effectively improved the quality and motivation of students' writing, as shown in Figure 6. Figure 6 Line graph of students' line quality and motivation under this model Facing the era of mass data, big data technology emerged and was gradually promoted according to the requirements of the times. When dealing with extensive data, big data has advantages that inherited information processing methods cannot match. It has the advantages of fast computing, fast information extraction and accurate feature marking. It is urgent to master the new era of university students and understand extensive data analysis methods, data processing operations, and big data predictive analysis. However, the restructuring of the university education system as a whole is at a standstill. Therefore, university students should learn about it by reading materials [14]. The creation of articles also covers various aspects; there are skills in using words and controlling the written Language. Traditional articles are often limited to a limited amount of information (pictures, texts, cartoons, news facts, Etc.), and the big do-it-yourself tower of writing covers all levels of social and productive life and everything. In addition, because of the vast data for support, it is objectivity and persuasive power are far better than the previous article and more in line with reality. In my opinion, college students in the age of information technology should have the ability to grasp and analyze the use of big data processing and then combine their literary skills and the use of writing skills to write good essays in the accompanying creative essays in order to write the ideas and rhythms of expression within the text and improve their literary skills [15]. The level of economic development differs from region to region. The corresponding educational system is constructed differently, and first-tier cities, which have the best teachers and educational resources, are at the forefront of constructing a new educational system. Depending on each region, there is a significant difference in the number of high-quality creative writing accompanying college students, as shown in Figure 7. Figure. 7 Map of sample data sources Big data can play a great potential in the new era. We can make intelligent analyses of learning based on big data, build a visual teaching platform through the network operation and maintenance system, use massive data for in-depth decision learning (predictive analysis through big data), and so on, which will help teachers build a diversified teaching system and have a profound impact in writing teaching We begin with predictive analytics using big data. We first use big data predictive analytics to build models to analyze large amounts of data and then use the learning results to make predictive judgments about the development of the matter. With the amount of data, this result will be very close to the actual value (of course, it is impossible to achieve 100% accuracy). Based on the above analysis, college students can control a stable and correct direction when writing the accompanying essay and make the essay come alive through the diversity of data.
In the teaching practice of teachers, they are usually troubled by these problems: they do not have a sufficient grasp of students' learning, they have less interaction with students in the teaching process, and it is difficult for teachers to analyze further and utilize the various learning situation data they have, and so on. In these situations, we can use predictive analytics based on big data to overcome these problems. We can use predictive analytics based on big data to carry out better teaching practices [16]. In digital writing teaching practices, students have reported excellent results in the face of personalized and customized learning experiences. In the case of digital writing instruction, students have reported excellent results with personalized and customized learning experiences and predictive analysis of big data based on the "information gain" measure of decision trees. We also integrate the teaching of such analytical skills into the teaching practice and cultivate the ability of predictive analysis based on big data for college students so that they can do high-quality creative writing and better adapt to the standards of the new era and make a summary contribution to the construction of a solid socialist country [17]. In digital teaching practice, we have established a teaching feedback mechanism based on big data, which objectively improves the effectiveness of composition teaching to a certain extent and provides an example for high school teachers in composition teaching. First, teachers, through their keen perception of current social issues, independently publish writing materials by collecting preliminary information, and students write essay ideas based on this material. After receiving the independent writing materials from the teacher, students look for writing ideas by understanding the materials. Each of us chooses the ideas we like or are interested in, and then there is a group division of labor in a collection of material data. This step requires students to receive material data independently, which is the first step to developing the ability to collect information with big data. In teaching creative writing with text in a high-quality context for college students based on the predictive analysis ability of big data, refined teaching is conducted to develop the following aspects of college student's abilities, as shown in Figure 8. Figure. 8 Radar chart of college students' quality Based on the theoretical part of this paper, teachers should focus on developing students' competencies in these areas in their teaching practice.
(1) Motivation Motivation is the premise of high-quality writing for college students. Without motivation, there is no subjective motivation; however, subjective motivation guarantees high-quality completion of the accompanying creative writing, and only by cultivating students' subjective motivation can they better complete the accompanying creative writing. In teaching practice, teachers should improve the quality and efficiency of writing by cultivating students' motivation. Only based on the content that interests students, students will take the Xu • Li 8 initiative to learn more and complete the writing practice assigned by teachers with their hearts.
(2) Autonomy Self-directed learning is the core ability of comprehensive training of college students, and it is also the type of ability we pay the most attention to under the modern education system. Students with such ability can adapt and face well in the face of society and later working life and have a solid resistance to stress. They can also face various problems in the practice of writing, teaching independently in high quality accompanying creative writing and can complete their writing tasks on time, with quality and quantity. Autonomy is a very critical type of ability.
(3) Analytical ability Predictive analysis based on big data also includes processing and analyzing massive data. Creative writing with text covers various aspects, including the skill of using words and the mastery of Language for writing narratives. Traditional writing with text is often limited to a limited amount of information (pictures, text, comics, news facts, Etc.), while writing with big data covers all dimensions of social production life and is allencompassing. Moreover, the description of affairs is based on massive data support; objectivity and persuasive power are much stronger than traditional writing and closer to reality. (4) Problem-solving ability The purpose of our teaching practice is not only to cultivate new-age college students with good literacy but also to cultivate students' problem-solving ability, which will accompany students' work and study for a lifetime, and is the ultimate purpose of our creative writing. Only when students' problem-solving ability is cultivated can we truly achieve the purpose of writing teaching practice [18]. (5) Literary literacy Literary literacy is not only the level of elegance of a person's literary level but also the vision and perspective of the person's view of the world, a way of thinking about the world that can help us better understand and transform the world [19][20][21]. Understanding history, reading literature and tasting philosophy are all ways to improve literary literacy, and teachers should focus on cultivating students' literary literacy in their writing-teaching practice [22].

Conclusion
With the advent of the digital age, big data has become a vital link in various fields. As a more indispensable component in digital education, it can cover students' learning progress, learning time, practice results, course information, classroom interaction information, Etc., genuinely covering the entire education curriculum for a one-stop learning experience. Big data drives China's education to advance continuously towards scientific, systematic, informational and counterpart. In recent years, big data has been applied to more and more educational fields, and predictive analysis based on big data has become one of the increasingly essential abilities of modern college students through all stages of learning. With the development of the times, college students write in a wide range of time and space. In such a context, the writing mode of college students has also required changes to accommodate the era of big data better. Big data predictive analysis has become an essential part of text production. One can write forward-looking essays only by being good at mastering and using big data for analysis. In order to further improve the composition skills of our college students, it is essential to integrate the systematic learning of big data into the daily teaching and life of the university. Writing is the most essential quality requirement for our college students today. Since human beings have written, writing has become a medium and an essential carrier for observing the world, describing the world and feeling the world. It is the core development requirement for college students under the requirements of the new era. College students who are proficient in writing use writing to record their lives on the one hand and express their perceptions and feelings on the other. This can not only improve the quality of humanities but also teach a way to find beautiful details in life, which can effectively release the pressure in the usual life and record the state of mood at this moment with the method of writing. Facing fastpaced study and life is a good way to regulate. Moreover, writing can cultivate students' imagination and creativity, which are the new comprehensive quality requirements for students under the new social development background, and for language teachers in colleges and universities, improving the ability of predictive analysis under big data and then integrating it into teaching practice to cultivate students with high-quality creative writing ability based on predictive analysis of big data is the current It is also the goal of universities to cultivate talents. Teachers should also put forward improvement opinions based on predictive analysis of big data according to the problems exposed by students in writing practice so that they can be targeted in the face of various problems and promote the steady progress of college writing teaching, forming a positive feedback adjustment mechanism of teaching-feedback-improvement. The positive feedback adjustment mechanism of teachingfeedback-improvement. To improve college students' high-quality creative writing ability based on predictive analysis of big data, we must develop students' active, independent, and analytical abilities, and also cultivate students' independent problemsolving qualities in teaching practice again, and then develop students' literary literacy through teaching practice based on the above, to cultivate high-quality college students for the new era.